at (ll«'a»«'iii'ivgiil |DIB«fl!||}ltaiNI iii«B«««gi i«iiii«»,:iivift«fli'» {W:»ftif»»: ft^aitf.ii'CMS^' • «ll9»««ftftfll t 3:.{|.pi.«ftft:p'(ii#':. lists'; •J pRflNKLiM Institute Library FHIL/lbELFHId Class to Book . IZ.S.3. Accession I.QJ..3B Article V. — The Library shall be divided into two classes ; the first ^ ^ duced, strain upon the warp is increased. The writer has examined many tappets made by different loom-makers, but has found few with a dwell exceeding \ ^ pick. The majority of tappets now in use in Lancashire for the production of fabrics with a good cover have only \ of a pick for dwell. Cover can be obtained by other means, but that Avill be dealt with at a later stage. The last of the points named refers to the treadle bowl employed to elevate or depress a tappet treadle without setting up undue friction at the points of con- tact. Generally speaking, a large roller works better and gives a steadier motion than a small one, but beyond this its diameter is of little importance, provided the incline on the tappet surface is not rendered too steep by the use of a large roller. The diameters in common use vary from If' to the smaller ones being used with tappets having the greatest number of picks to a repeat. A tappet- Ill SHEDDING OR DIVIDING THE WARP 39 maker must carefully consider what effect an antifriction roller has upon the motion of healcls, or they will rise and fall at times, and speeds, other than those intended. A tappet has its surface formed into a series of elevated and depressed parts, more or less resembling teeth in a wheel. It is self-evident that a 'point could be made to 2 Fin. 22. follow the most minute variations of such a surface ; but it is impossible to use a point, for that has position but not magnitude ; therefore an antifriction roller, which has its centre where its point should be, is employed, and, as a consequence, it is capable of filling, or partly filling, the space between two teeth, and also of touching two parts of the tappet surface at one time. Whatever space exists between the extreme points so touched is lost to the pause 40 MECHANISM OF WE A VING PART of the healds, and one begins to rise or fall l)efore another. The only plan by which this difficulty can be overcome is to make the roller centre move exactly as the healds are required to move, and to do this the size of tappet must be reduced to allow for the size of roller. If a tappet is -r, constructed for one roller it " ' will never work as satisf ac- ^ ' " V torily with one of any other dimensions. Enough has been said of the requirements of a tappet to enable the student to proceed with its con- struction. Fig. 22 is a negative tappet constructed to weave plain cloth. Fig. 23 one for a three-thread twill, one down, two up. Figs. 24 and 25 are four - thread twills, one down and three up, and two up and two down for each pick. Fig. 28 is a positive tappet for the 6-pick pattern which accom- panies it. Assume the plain cloth tappet to be 1^' from centre to thinnest part, to have a lift of a dwell of J of a pick, and a treadle bowl 3'' in diameter. Describe circle a (Fig. 22) equal in radius to the dis- tance between centre of tappet shaft and centre of treadle bowl when bowl is touching thinnest part of tappet = 1^4- radius of bowl I J tappet 3" = 2J"-f 3" Fig. 23. 2f", add to that the lift of 5|", and describe a second circle h. Ill SHEDDING OR DIVIDING THE WARP 41 When in action the treadle bowl constantly works on or between these lines. Divide the circles into two equal parts, because the pattern has 2 picks to a repeat, by lines 1, 4. Subdivide each space into three equal parts, and draw radial lines 1, 2, 3, 4, 5, 6. Space 1, 3, equals | of a pick, and is to be used for moving a shaft for the first pick. Space 3, 4, equals J of a pick, allowed for dwell. Divide spaces 1, 3, for the first, and 4, 6, for the second picks, also by radial lines, into any number of equal parts, say 6 each. De- scribe a semicircle c upon any one of the radial lines, equal in diameter to the space between circles a, ^, and touching both ; divide its periphery into equal parts corresponding in number to those in division 1, 3 — namely 6. From each point of inter- section on the periphery of the semicircle drop a line d per- pendicular to its diameter line; the latter will then be divided into unequal spaces for the purpose of imparting an unequal velocity to the heald. Through each point on the diameter line describe a circle concentric with a, Consider each point, where a radial line is cut by a curved one, as the centre of the treadle bowl at diff'erent parts of the lift, then from those points, beginning on the outer circle, and taking the others in rotation, describe circles equal in 42 MECHANISM OF WEAVING TART diameter to the treadle ])owl, and the construction lines are completed. To find the shape of tappet, draw a line touching the periphery of each bowl circle, and where the heald is to be stationary, describe an arc from the tappet centre. What remains to be done is merely a question of removing surplus metal, as clearly shown in Figs. 22, 23, 24, and 25. A tappet treadle swings on its fulcrum pin, and causes the points of contact between bowl and tappet, relatively to the centre of the latter, to change continually ; but the movement required is a rise and fall in a straight line, hence to approximate to the best motion, the treadle fulcrum and bowl centres should be in the same horizontal plane at half the lift ; and bowl and tappet centres should B Fio. 25. Ill SHEDDING OR DIVIDING THE WARP 43 then coincide, or the treadle will neutralise to some extent the lift aimed at in constructing the curved surfaces of the plates. Treadles must be in contact with tappet at every point of its revolution, or the healds will have a jerky movement. In order to place warp level where the shuttle enters it, a back shaft must rise higher and sink lower than a front one ; but this cannot be accomplished if all the tappet plates are of one size and the treadle fulcrum pin is at the back of loom, for the connection to back shaft is made at a point nearer the fulcrum than for a front shaft, hence its movement would be less instead of greater ; so it is usual to give an increased lift to a back plate of from \' to but on an average it equals more than a front one. If a treadle fulcrum is at the front of a loom the back shaft connection is further from it than that of the front shaft, for which reason all plates are of one size, and the difference in leverage gives the difference in lift. If the stepped rollers, called cones, are properly con- structed, heald straps will be kept at the same tension at every part of the tappet's revolution ; they should be made so each will wind on or off exactly as much strap as its tappet plate gives out or takes up. Jamieson's Tappet A negative tappet introduced by Jamieson is much used in certain districts of Lancashire. It is placed under the healds with its axis at right angles to and coinciding with that of the bottom shaft rt. Fig. 26. At its rear end the tappet shaft h has a large bevel wheel c to gear into one of two small bevel pinions 6, e', slided upon shaft (X, with their teeth facing, for the purpose of driving c in either direction. PART III SHEDDING OR DIVIDING THE WARP 45 A tappet wheel d is keyed upon V to gear with c , and contains a series of bolt holes concentric with h' and equi- distant from each other ; the number of holes equals the picks to the round of the tappet. Bolts are passed through the holes, and a plate / similarly drilled is dropped over them ; its periphery has a series of slight curves corresponding with the holes. One of its faces has a number of slotted recesses, through the centre of each a bolt passes, and into which an ear of an interchange- able plate g is dropped until a portion of the latter rests upon the outside of /. Each plate g sinks a shaft, and an uncovered curve upon / permits one to be elevated. It is therefore simply a question of reading from a tie- up, as for a Woodcroft, and placing sinkers in position on as many plates / as there are shafts, then bolting the whole firmly together. Treadles % are placed below the tappet, and each has an antifriction bowl riveted to it to work against the projec- tions on / and pull the shafts down, but springs attached to a frame above the Idom lift them. Four differently shaped sinkers are sometimes used — one to sink a shaft for a single pick, two serve as right and left hand sinkers when a shaft is to remain down for two or more picks, a fourth gives the dwell, or supports a treadle bowl as it rolls from one rising part to another. The sinkers are liable to get loose, and the bowls A to slip between two plates ; beyond this it is a good and use- ful tappet. The Barrel Tappet Another negative tappet known as the barrel motion is also a favourite in some manufacturing centres. It is fixed above and at right angles to the heald shafts. An upright FART III SHEDDING OR DIVIDING THE WARP 47 shaft a (Fig. 27) and bevel wheels h are employed to drive it, either from the bottom, or crank shaft. Some makers resort to the objectionable practice of placing the upright in such a position that it passes through the warp, instead of fixing it to the end framing, and driving tappet d through a short horizontal shaft c. This arrangement necessitates the use of two additional bevels, but the work- ing parts of a loom are left unobstructed. The tappet consists of a series of solid plates bolted together or cast in one piece ; if the former, a hole is made in every plate exactly in the centre of each pick, so that by turning similar plates into different positions, one with relation to the other, many different patterns can be woven. If the tappet is all cast in one piece fewer changes are possible ; but, on the other hand, it is solid, and there is no risk of the bolts working loose. Treadles rest above the tappet, and springs or other appliances pull the shafts down. When arranged for bordered fabrics, such as handker- chiefs, towels, cross stripes, or other fabrics requiring two patterns in the same piece, a barrel is furnished with twice as many plates as there are shafts to be moved, and the treadles act on alternate plates. By shifting a lever the tappet makes a lateral movement and slides the other series of plates under the bowls to weave the second pattern. Positive Tappets A positive tappet (see Fig. 28) is constructed on similar lines to a negative one, except in so far as a minimum space of must be provided between the outer and inner flanges above that required for the treadle bowl, to leave 48 MECHANISM OF WE A VI NG PART room for its easy working, and reduce the chances of it be- coming blocked by an accumulation of dirt. In setting out one of these tappets it is a common practice to add ^' to the lift, and after finding the points of intersection between radial and concentric lines, as ex- plained on p. 40, to describe circles round each, of a diameter exceeding that of the treadle bowl by ^\ then to draw lines touching their peripheries at two points, which trace the inner and outer flanges. After that it only remains to add the required thickness and depth to give the necessary strength and surface for the bowl to run against. An inward curve on the outside flange equals a lifted shaft, and an outward curve on the inner flange a depressed one. Ill SHEDDING OR DIVIDING THE WARP 49 Woodcroft's Tappet In 1838 Woodcroft patented a positive tappet that is still largely used for heavy work, and also where eight or more picks to the pattern are required. 1 7 S Fig. 29. It consists of corresponding sections cast with an elevator or depressor on each, which, when placed together, form the entire plate. By changing the relative positions and thus forming new combinations of sections, patterns may be varied at small cost, provided the picks in the new E MECHANISM OF WE A VTNG PART pattern equal the number of sections in a full plate, or that two or more full patterns are contained on a plate. For example, in a tappet known as "16 to the round" a section equals -^^ of a plate, therefore any pattern of 2, 4, 8, or 16 picks can be made with the same sections if suitably arranged, but if a pattern has any other number of picks, new sections must be used. In Fig. 29 a shows the form of an elevator, and h a de- pressor. Dark squares on the slip of design paper (Fig. 30) indicate a rising shaft, and blank squares a sinking shaft. As eight squares are used, eight to the round sections are 4 necessary, and they must be placed to correspond g with the position of the tappet and the direction g of its motion. If fixed at the right-hand end of ^ loom, to revolve in an opposite direction to the ^ crank shaft, place the tappet wheel before you, with g a number of bolts previously passed through the bolt holes provided. Lay an elevator in position g (number 1 on the design paper) and follow with number 2 a depressor, on the left side of number L Continue reading from the design and laying the sections until all are in position ; then put the outer flange of a ring (Fig. 31) between projections c, and its inner flange below d to lock all together. Proceed to form other circles of sections from similar slips of design paper, one for each heald shaft employed, always beginning immediately above number 1 of the first plate. Bolt all together and the tappet is ready for the loom. When a tappet revolves in the same direction as the crank shaft, lay the second section on the right of number 1. A tappet fixed at the left-hand end of loom, to revolve ITI SHEDDING OR DIVIDING THE WARP 51 in an opposite direction to the crank shaft, requires a ring to be first placed upon a blank plate on the tappet wheel, and the second section laid to the right of the first, and all face side down, or the treadle bowls cannot be made to run in the groove between two plates. J. Fig. 31. If turned in the same direction as cranks, lay the second section on the left of number 1. As mentioned on p. 21, Woodcroft's tappet is now made to produce open shedding, but the parts are some- what complicated, for, in place of using duplicates of two 52 MECHANISM OF WE A VING part sections for each plate, eight distinct sections are necessary — namely, ordinary risers and sinkers, right and left hand risers, right and left hand sinkers, riser dwells and sinker dwells. Each section is clearly shown in Fig. 32 : number 1 is a 1 5 Fio. 32. riser \ 2, a sinker ; 3, a left-hand riser ; 4, a riser dwell ; 5, a right-hand riser ; 6, a left-hand sinker ; 7, a sinker dwell ; and 8, a right-hand sinker. With the above exceptions plates are laid in the manner described on the preceding pages. Ill SHEDDING OR DIVIDING THE WARP 53 In Fig. 33 an antifriction roller /, carried by a treadle centred at is pushed up and down by the sections and imparts an oscillating motion to the outer end of g. Straps and cords pass up and down from treadle g to top and bottom jacks li^ i, which are fixed in the loom at right angles to g, but in the sketch are turned from their working posi- o i 1 — 1- ^ — 1 H ID Fi(}. 33. tions through an angle of 90°, in order to open them out and show the working more distinctly. The dotted line o indicates the hinge on which the parts are turned. The fulcrum 2)ins of h, i, are shown at j, k ; other cords, /, pass down and up to the top and bottom heald shafts ; therefore if treadle g is raised shaft n is depressed, or if g is depressed n is raised. A Woodcroft tappet is not so firm as one made from 54 MECHANISM OF WE A VING PART solid plates, and there is a constant tendency for the treadles to leave the tappet in cases where warp threads are too few to put considerable strain upon the shafts. To prevent which, springs and w^eights are attached near the fulcrum pin of the treadles to hold them down. Oscillating Tappet Chains of various kinds have been used for more than fifty years as shedding motions. Clarke's invention dates from 1840 ; Knowles's from 1849. Several others have also been patented at different times, of which Nuttall's is one of the best ; his chain is composed of rollers a and collars & (Fig. 34) pushed upon long spindles c, the latter are connected at each end by flat links d. A roller lifts a shaft, a collar sinks one, and at the same time holds the rollers in their proper places. The chain is divided into two parts — one containing all the odd picks, and the other all the even picks. Both are passed round eight-sided barrels that are free to turn in bearings attached to the tappet. A flat piece of metal /, weighted at its outer end, forming part of an elbow lever ^, rests upon the chain, and is elevated by a roller, but falls by gravity if a collar is upon the uppermost side of barrel e ; this imparts a corresponding motion to which has a projection h that enters a slot in tappet plate i ; the latter, being fulcrumed at y, is lifted by a roller into the position shown at i, on the right of the drawing, but a collar leaves it, as at i on the left. The tappet is made to oscillate in the following manner : — Pinion 1 (Fig. 35), on the crank shaft drives carrier wheel 2 round stud and it gears with slide wheel 3 ; the teeth in the latter must be in the proportion of tw^o to one in the pinion. A boss compounds wheel 3 with eccentric 4, they revolve Ill SHEDDING OR DIVIDING THE WARP 55 upon stud and by means of eccentric strap, and arm a connection is made with a lever o centred at that carries a second arm which is fastened to pin r on the tappet. As 4 goes round the tappet rocks, if to the left, treadle bowl 5 (Fig. 34) will run along the under side of i, depress a treadle t vibrating on pin / in the loom framing and lift a heald shaft ; but if the tappet rocks to the right, s will roll Fig. 34. along the upper side of i, elevate ^, and depress a shaft, for heald connections (Fig. 14) are similar to those of a Wood- croft tappet. Bowl 5 can run above or below the plates i, any number of times in succession, its position depending entirely upon the construction of the chains ; the latter must be rotated to place the picks progressively in the fabric ; this is accomplished by slide and star wheels ; thus the slide wheel 3 (Fig. 35) has a circular flange 5 shown broken at the top where stud 6 is inserted. Star wheels ^, ^, are re- spectively keyed upon each chain barrel, and as slide 3 56 MECHANISM OF WEA VING part revolves, it carries stud 6 alternately into a notch of stars Fio. 35. 6, e : and every time this takes place, a barrel is turned \ of a revolution. The advantages of this motion are : that it is positive ; that long patterns of from 80 to 100 picks can be woven Ill SHEDDING OR DIVIDING THE WARP 57 if the tappet centre is placed above the crank shaft, because the chain is divided into two parts ; that the same bowls, collars, and pins can be used for all patterns by simply relaying the chain, but it is more suitable for heavy than light fabrics. Nuttall's Chain Another positive chain tappet, patented by James Nuttall, is also much used in districts where heavy goods Fig. 36. are made. All its parts are contained in a movable stand that can be taken to the end of any loom, fixed to the floor, and driven by a pinion from the crank shaft. It consists of two parallel chain barrels, a, h (Fig. 36), which are provided with suitable bearings, and employed to rotate chains, c, t/, at the required rate through spur wheels, /, ^, keyed upon the barrel shafts ; /, g are connected by a small carrier, A, that turns both in the same direction. The chains 58 MECHANISM OF WE A VING PART are composed of rollers and collars, not unlike those on the oscillating tappet, but here a collar on one barrel must be opposite a roller on the other, for both chains move at the same time, and act upon the same straight levers, one of which is shown at e resting upon them. This lever is ful- crumed midway between the centres of a, and its rear end projects slightly beyond &, but its forward end is long enough to permit of a strap attachment for cording to the upper and lower jacks. The tappet is not so well adapted for long patterns as the oscillator, for practically a link of each chain is required for a single pick ; the first to elevate the rear ends of e to lift shafts, and the second to elevate their forward ends to depress shafts ; but it is positive in action and simple in construction. Tappet Driving Tappets are driven, directly or indirectly, from the crank shaft by spur wheels, and whenever convenient, it is best to have a large one on the tappet and a smaller one on the crank shaft, as then the proportion equals the revolu- tions of the two shafts ; thus a 4-picked tappet requires a wheel with four times as many teeth as that on the crank shaft ; or put in another form — if a tappet wheel contain- ing 120 teeth is to drive a tappet at the loom end, having 2 picks to the round, 120-^2 = 60 teeth in the driving wheel. If 5 picks to the round, 120-^5 = 24 teeth. But it frequently happens that two wheels will not answer, because both shafts are fixed, and the wheels, if not of a proper size, will not gear ; in such cases a carrier or carriers are employed (see taking up calculations. Part XIX.). A single carrier will cause a tappet to revolve in the opposite in SHEDDING Ok DIVIDING THE WARP 59 direction to one without, or with two carriers. Fig. 37 shows the tappets placed under the healds, but driven from a short shaft a, to which motion is given by crank wheel ^, bottom shaft wheel c', a second wheel d on the same shaft, a carrier ^, and tappet wheel /. If Z/ = 40, f ■ I \N\, \V Fig. 37, c = 80, 6? ==20, and /=40 teeth respectively, the revolu- tions of a will be to those of crank shaft as 40 X 20 80 X 40 ' 1 :4. Plain cloth tappets are generally keyed to the bottom shaft, because picks to the round = 2, and revolutions of shaft = I those of cranks. It is not unusual to employ compounded intermediate wheels to drive tappets having a large number of picks to the round ; the intermediates work loosely on a stud, the larger one gears into the crank shaft Avheel, and the smaller 6o MECHANISM OF WE A VING PART one into the tappet wheel. There are thus two driving and two driven wheels. As a rule the tappet wheel is bolted to the tappet and cannot be changed, but the three remaining can be found by dividing the tappet wheel by any number that will not leave a fraction. The divisor will give the teeth for one pinion, the quotient will give the other ; the picks to the round give the first driver, and the tappet wheel is the second ; thus, in a 16-picked tappet, with a wheel of 180 teeth — 180 -M5 = 12 . • . 12 into 16 (the picks) and 15 into 180. 15x12 1 for proof = — or 1 6 revolutions to 1 . ^ 16x180 16 , . 12x20 1 Ai^am, 180-M2-20.-. = — . ' 16x180 16 But assuming that a wheel with fewer than 20 teeth cannot be used, any multiple of the numbers obtained will answer, as 15x2, 12x2; but the tappet wheel being fixed, 16 must be multiplied by 4, because the teeth in driving wheels have been increased fourfold, and to maintain the ratio, teeth in driven must be proportionately increased. ^a A T.^ , 30 x 24 1 .-.16x4 = 64. i or proof = -— -• ^ 64 x 180 16 Both intermediate wheels can be found if crank pinion and tappet wheels are given, by rule — Picks required x pinion Teeth in tajDpet wheel Example. — If 20 pinion and 120 tappet wheels are used to drive a tappet of 12 to the round, what intermediates are required 1 20x12 2 1 T . ^ . X. . . . • . any wheels m the ratio ot 2 to 1. 120 1 IV OVER AND UNDER MOTIONS 6i Changing 'one wheel is often sufficient; it may be a driver or a driven, but the rules are — Driving wheels x picks to the round { teeth required for a Driven wheel I driven wheel ; and Driven wheels Driver x picks to the round : teeth in a driving wheel. Scroll tappets are rapidly becoming favourites in those centres where coloured shirtings are manufactured, and they appear to be well adapted to the production of narrow stripes and chain effects. In such cases scrolls merely govern a few threads that rise for three, four, or more picks in succession, and then remain depressed for an equal number of picks, whilst all remaining warp is actuated by tappets of the usual form. The scroll, as a shedding motion, is by no means a recent invention, nor does it appear likely to be ever employed for general purposes ; it is similar in construction to a picking scroll (see Part X. Fig. 171), but the half -moon is attached to a tappet treadle. Pendulum and other tappets are frequently met with, most of w^hich are capable of producing good work, but want of space precludes a detailed description. PART lY OVER AND UNDER MOTIONS All negative shedding motions require additional parts to reverse the direction of their action. Such appliances are known as over or under motions. These names seem to imply two classes of mechanism, but in reality the dis- 62 MECHANISM OF WE A VING PART tinction is merely one of position. If a tappet is placed under the healds, an over motion is used, but if it is placed, over them, or at one end of the loom, an under motion becomes necessary. Eeversing motions may, however, be grouped as a, single acting ; compound acting. In the former, each part is only capable of exerting force upon a single shaft ; but in the latter, each part can be made to act simultaneously on all the shafts in a set except one. Before proceeding to describe these motions, a short time may be well spent on an examination of the nature of the work to be done by them. It has been previously shown that a negative tappet can move a shaft in one direction without outside aid ; this being the case, the action of an over or under motion should begin immediately that of a tappet ceases, and con- tinue with increasing force until the termination of the shaft's journey is reached, where its maximum power must be exerted and maintained until the tappet is again brought into use. During the period of the latter's action little or no force is needed to oppose it, for if a moving shaft is kept steady, nothing more is required, and any additional strain is power wasted. These remarks are especially applicable to centre and open shedding, where in both cases the tension upon the warp threads exerts its greatest power to pull the shafts down or up to the closed shed line at a time when the reversing gear has to resist such strain unaided by other parts of the shedding motion. Single Acting Motions Single acting motions are generally composed of dead weights, spiral springs, or elastic cords (see Figs. 38, 39), which are attached to the heald shafts, but none of them IV OVER AND UNDER MOTIONS 63 are mechanically correct. A weight exerts equal power at every point, no matter whether the tappet is acting or has ceased to act. Undesirable as this is, weights nevertheless more nearly approach our ideal than springs, and they are invariably emploj^ed for Jacquard shedding, but have not met with much favour when applied to shaft work. Their defects are obvious ; if rapid reciprocating motion is imparted to weights more or less free at one end, an unsteady, swinging action is fre- quently set up, and even when swinging is prevented by confining the weights in grids, considerable friction results. Another objection based on a well-known law in mechanics has been raised against the use of dead weights, viz. the shorter the time during which a body is allowed to fall, the slower its movement. Those who object to weights on this account say that the rapidity of motion required, and the short space through which they have to move, sometimes result in changes being completed in one part of a harness before they are completed in another part. If what has been said respecting the functions of reversing motions is accepted as the true principles which ought to govern their action, it will be an easy matter to show that a spiral spring is one of the most defective pieces of mechanism for the purpose that can well be imagined. A spring must be stretched until it exerts sufficient force to pull the warp threads it controls level with the top or bottom shed lines when the tappet is inoperative ; and 64 MECHANISM OF WE A VING part therefore the further a tappet moves a heald from that point, the more a spring is stretched, and the greater the force it exerts to oppose that of the tappet. On the latter ceasing to act, the former is exerting its maximum power IV OVER AND UNDER MOTIONS 65 to pull the heald shaft in the opposite direction, and this takes place when least power is needed. Immediately a reversion occurs the spring contracts, and its force decreases in proportion to that of its increase when stretched until a point is reached where its effectiveness is most needed, then it exerts its minimum force. A spring is thus seen to be strongest where it should be weakest, and weakest where it should be strongest. If a comparison of the power consumed by springs and dead weights is made, this point will be rendered clearer. On the assumption that a force equal to a weight of 10 lbs. is needed to hold a heald shaft stationary, a spiral spring must be stretched until it is capable of exerting that force. If it is further assumed that such stretch = 1'', and the movement of the heald = 4'', then as the force of a spring increases uniformly 10x5 = 50 lbs., the power exerted when it is not wanted, and 10 lbs. its effective force. Or put into units of inch pounds 50 2 = 25 lbs., the mean power, and 25x5 = 125 inch pounds. A dead weight of 10 lbs. exerts equal force at all points, . • . inch pounds = 10 X 4 = 40. It is thus seen that under the above conditions the work done during one movement of a shaft is with weights 40 inch pounds, and with springs 125 inch pounds. Notwithstanding the faults inherent in spiral springs, they are more extensively used in Lancashire for light and medium fabrics requiring 6 shafts or more than any other appliance, probably because they are readily fixed and easy to understand. Inventors have, however, endea- voured to introduce motions with a spring basis so arranged that as a shaft moves the full force of a spring will not be exerted upon it. The following illustrate two of the attempts made to solve this problem. : — F 66 MECHANISM OF WE A VI NG PART kSpring-easing Motions Messrs. Hahlo, Liebreich, & Hanson have a patent easing motion for allowing a shaft to move from the bottom to the top shed without stretching a spring an ¥iv,. 40. equal distance. In construction the motion is simple, and the parts are not liable to get out of order. They consist of lever a (Fig. 40) fulcrumed at to which the upper end of spiral c is attached by a wing nut and screw used to regulate the tension, or as shown at d. The lower end of c is fastened by strap c upon eccentric /, whence c passes IV OVER AND UNDER MOTIONS 67 to hook (j^ and connection with lever a is again made at point h. A heald shaft is connected to the outer end of a. There- fore, as a and g ascend together, the latter will pull strap from the thick side of eccentric / and wind it upon the thin side ; but point d moves down simultaneously, and thus prevents spring c from stretching in proportion to the winding of strap e upon eccentric / The surface of / and the position of d are so contrived that, for a movement of b" in a shaft, spring c will be stretched f ; this results in a considerable saving of power as compared with springs used in the ordinary way. Thus, on the former assump- tion of 10 lbs. equalling the weight required, and also a stretch of V\ an ordinary spring will exert a force of 5x10+10 = 60 lbs., but used as in the above motion equals a force of 10 xf of 10 = 17|^ lbs. Kenyon's Under-motion Kenyon's motion is equally simple, but it acts in a somewhat different manner upon the shafts. In Fig. 41 two chairs, tt, h, are placed back to back and bolted upon a rail c, on which bolts and drag nuts provide for moving the chairs nearer to or farther from each other in case spring d is too strong or too weak. Chains f are hooked into opposite ends of d^ and are respectively con- nected to the hooked parts of curved levers g, h ; the latter move partly round fulcrum pins and at Z;, I straps pass to a shaft m. The form of levers ^, h causes spring d to be elevated and depressed as its shaft rises and falls, and the distance so lifted equals a loss of stretch in spring d ; but that is not all, for when points of connection n, 0 are below centres 68 MECHANISM OF WE A VING PART of pins y, the principal stress is upon the shaft, but immediately that centre is passed, an increasing stress is transferred to the fulcrum pins, with the result that the actual stretch of spring does not represent the force applied to pull a heald down. For example, it was found that a lift of ^" stretched spring d 2|'', which, under the previously assumed conditions, would equal 2|xl0+10 or 37|- lbs. when at the highest point, and 10 lbs. when at the lowest. An actual test made with a spring balance gave the ratio of stress as V2\ lbs. for a closed, and 6^ lbs. for a shed opened 5". Taken inch by inch the stress varied as follows : — Closed shed, 12|^ lbs. ; first inch of lift, \\\ lbs. ; second inch, 8f lbs. ; third inch, 6f lbs. ; fourth inch, 5| lbs. ; and fifth inch, 6| lbs. Compound Motions The second class of reversing motions, known as stocks and bowls, are all based on the compensating principle : this consists in making a rising or falling shaft help to move another in the opposite direction. Whenever such a motion can be successfully applied, a given piece of work IV OVER AND UNDER MOTIONS 69 Avill be performed with a smaller consumption of power than with w^eights or ordinary spiral springs. An approximation to the relative consumption of power between weights and stocks and bowls can be obtained as follows : — Assume the closed line of warp to be midway between the highest and lowest points, and that the power required to move a shaft up or down from such position equals one unit of work, then the weight upon the shaft must equal one unit, or it would not occupy its proper position. In lifting a shaft from the lowest to the highest point one unit will be consumed in moving it to the centre — namely, lifting the weight — and two additional units will be used between the centre and top line — one in moving the weight, the other in moving the warp ; therefore three units of work will be taken from the engine for each upward motion of the shaft, but the weight will exert sufficient force to carry the warp back to its lowest point. With stocks and bowls working under similar conditions the tension of the warp is sufficient to lift and sink shafts which are coupled, to the closed position ; therefore, in moving one from bottom to top, and the other from top to bottom simultaneously, two units are required — one for the upward, and one for the downward movement, and two additional units must be used to take the shafts back to their starting-points. Four units of work have thus been taken from the engine, but in this case two shafts are moved to one in the former, consequently the approximate consumption of power with Aveights is six units to four units with stocks and bowls. Stocks and bowls consist of levers either circular or oblong in form, to which the shafts are connected, but they can only be employed when the pattern to be woven 70 MECHANISM OF WE A VING PART requires the same number of shafts lifting for every pick. In cases where say two shafts are lifted for one pick, and three for another, a single acting motion must be used. Shafts are sometimes connected in pairs to single levers, as seen in Figs. 42 and 43 ; the former arrangement is for plain, and the latter for twill weaving. This principle can be extended to work any even number of shafts, but IV OVER AND UNDER MOTIONS 71 half the number employed must go up and half down at each pick. In this form stocks and bowls are therefore^, inapplicable to a large number of patterns. But when levers are compounded, as in Figs. 44 to 53, any number of shafts can be carried up or down at one time, provided, as previously stated, the same number go up each pick. Figs. 44 and 45 are arranged for 4 and 8 shafts. In the 1 2 3 9 4 £ D n r — □ r C B ri 0 1 A Fk;. 44. first straight levers are employed. Lever a supports by means of connecting straps h, c, levers d, from each end of which a strap 1, 2, 3, or 4 passes to its respective shaft. Fig. 46 shows an arrangement of bowls for moving a similar number of shafts. A strap passes round bowl a, and supports bowls 1\g; strap 1, 2 passes round I, and is attached to shafts 1,2; strap 3, 4 passes round r, and is attached to shafts 3, 4. The remaining motions consist of a larger number of bowls connected to each other and to the shafts in a 72 MECHANISM OF WE A VI NC PART similar manner to the above, a further description of them being unnecessary. When an odd number of shafts have to be used more Fio. 45. Fin. 40. bowls are placed at one end of the first lever than at the other ; and in all such cases the fulcrum of the first lever must be placed to balance the weight — namely, the number of shafts attached to both ends. Thus in a 3-shaft motion the bottom lever is divided into three equal IV OVER AND UNDER MOTIONS 73 parts, and the fulcrum pin placed at the first division from the heavy end, as in Fig. 47, where a is the bottom lever, h the fulcrum pin, c a strap passing up to number 1 shaft, d a strap supporting lever e at its centre, and /, g straps con- nected to shafts 2, 3. When bowls are used throughout, the two lower ones are compounded by fastening both together and screwing a strap to the periphery of each. The diameter of one bowl 1 2 3 • C c E n r D B 1 r + 0 A Fig. 47. must be to that of the other in inverse proportion to the number of shafts each controls. Thus in Fig. 48, which is a 3 -shaft arrangement, the diameter of a is twice that of because strap 1, screwed to bowl is attached to shaft 1 only ; but strap 2, screwed to bowl controls through bowl and strap 3, shafts 2, 3. There are many modifications in the details of stocks and bowls, but by suitably combining a 2, 4, 6, or 8 shaft motion with one for 3 shafts, any odd number up to 74 MECHANISM OF WE A VING PART 11 shafts can be worked. The following combinations for 5 and 7 shafts will sufficiently explain the methods : — In Fig. 51a 3-shaft motion (Fig. 48) is attached to one end of a straight lever a, and a 2 -shaft motion (Fig. 42) to the opposite end ; a is divided into five equal parts, and the fulcrum pin placed two divisions away from the 3-shaft motion, to balance the weight. In Fig. 53 the 3-shaft motion is re- tained, but that for 2 shafts is replaced by one for 4 shafts (Fig. 46). Lever is now divided into seven parts, and the pin placed at the third line from the heavy end. In order to prevent stocks and bowls from giving a side pull to the healds it is usual to employ two sets, each one placed at the same distance from the extremities of the shafts \ and to avoid the straps twisting between the rollers and shafts, the axes of some upper rollers are frequently at right angles to those below. In Fig. 51 a set is shown in work- ing position wath the requisite connec- tions to the shafts. The upper rollers are also turned to bring each strap immediately under the shaft to which it is attached. Owing to the multiplicity of levers in some sets of stocks and bowls it is not always an easy matter to trace the movement of each roller when a shed is forming, but a little careful analysis will be sufficient to demonstrate it. Something depends upon the character of the shedding Fio. 48. IV OVER AND UNDER MOTIONS 75 motion. If, for instance, a Jamieson tappet, a single lift, or a centre shed dobby is used (see p. 43), all warp is placed iu one line after each passage of a shuttle, but the position Fio, 49. Fifj. 50. of that line varies with the number of shafts moved up or down at one time. Its height above the lowest point of a shed can be found by employing the following rule : — Depth of shed x shafts lifted for one pick Shafts in the set. 76 MECHANISM OF WEAVING PART Examiple. — If a shed Z" deep is made for weaving a five- thread twill, three up and two down, the dosed point will be 3x3 9 above the bottom shed line ; bub if the same size of shed 1 LUJ I A ] \ ^ t (®) 5 + 4 + Fig. 51. is made for a four up and one down twill pattern, it 3x4 12 = = — = 2-4" 5 5 from the bottom. For proof refer to Fig. 54, in which a is the pattern, h three divisions ruled to represent inches IV OVER AND UNDER MOTIONS 11 in depth of shed, each being subdivided into five parts to correspond with the number of shafts employed. Vertical divisions 1, 2, 3, 4, 5 contain respectively five dots c, that Fig. 52. Fig. 53. show warp level for five successive picks ; d is the under- motion used to connect the shafts. Assume shafts 1, 2 are lifted *. • -^-j^" = Z" - 2-^^, shafts 3, 4, 5 will sink in the proportion of 3:2:: : (see bottom lever e), tV + tV = 1^ the distance below 1, 2 ; 12 3 4 5 B Fig. 54, PART V DOBBY SHEDDING 79 and lifting 3, 4 from that position to the level of 1, 2, will pull down 5 in proportion to the surfaces of compound rollers / — namely, 1 : 2 : : : 2'', . *. when shaft 5 is on the bottom shed line, 1, 2, 3, 4 are on the top line. The move- ment of shafts for picks 2, 3, 4, 5 is traced on the figure^ but it must be borne in mind that it is simultaneous, and not consecutive for each pick. An open shedding motion acts somewhat differently, for all the shafts can never be at one height at the same instant. In tying them up, some are corded level with the top, others with the bottom sheds, in which positions they remain stationary until the pattern renders changes neces- sary. The same number of shafts rise and fall every time a shed is formed, and those remaining are locked either at the top or bottom by the shedding motion. If the last pattern is taken to illustrate this system, we find shafts 1, 2, 3, 4 up and 5 down for the first pick ; for the second, 5 goes up and 4 down 3''. Under-motion d shows that 3'' of strap will be unwound from large pulley /, and \y' wound upon the small one compounded with it. Shaft 3 being fixed, and the centre of roller g pulled down \y\ it follows that shaft 4 will be depressed — namely, by the fall of ^, and \)^' by increasing the length of strap 3. A similar movement can be traced amongst the rollers and straps when any pair of shafts are changed. PART V DOBBY SHEDDING If a pattern is beyond the range of a tappet cither in the number of shafts to be manipulated or in the picks to 8o MECHANISM OF WE A VI NG PART a repeat of the pattern, and is at the same time too small to be economically produced by a Jacquard, a machine specially constructed to control shafts is employed; it is known as Dobby, Witch, Wizard, and Index in different manufacturing centres, and the number of shafts it is capable of working is frequently mentioned with the machine, as a 16-shaft, or a 40-shaft dobby. In the immediate neighbourhood of Manchester large dobbies are not in extensive use ; from 12 to 16 shafts include by far the greatest proportion of these machines, although in a few instances as many as 24 shafts are employed. In the Dhooty trade 40 jack dobbies are common, but they are used to work a mail mounting, and for all purposes can be considered to do the work of a small Jacquard. In the Worsted industry 36 jack dobbies are in everyday use, and in exceptional cases as many as 70 jacks are employed. The last-named dobbies are usually positive in action, and are therefore preferable to a Jacquard for heavy cloth, firm shedding, and general good working. In places where patterns are often changed, a dobby is used for fabrics quite within the range of a tappet, as it offers greater facilities for producing a variety of effects than a tappet. Still it must be borne in mind that healds never work better, never last longer, or give greater satis- faction, than when actuated by a properly constructed tappet. It is probable that the dobby dates from a time anterior to that of the Jacquard. Certainly, analogous machines were applied to the hand-loom for weaving patterns beyond the range of treadles before Jacquard's invention reached this country, hence they can scarcely be said to owe their existence to the larger machine. As now made they are single-acting, double-acting. V DOBB V SHEDDING Si negative, and positive ; they form closed, centre, semi-open, and open sheds. In short, the variety is so great that it goes a long way towards proving that none are quite satisfactory, for wherever a piece of mechanism is found to be pre-eminently suitable, it soon displaces all else intended for the same purpose. In dobbies, one is too slow in action, another is only adapted for light fabrics, a third is too complicated in construction, a fourth results in excessive wear and tear, and so on. Beginning with the single lift negative dobby, we find a framing a (Fig. 55) placed over the centre of the shafts, which contains a number of vertical pieces 5, called hooks, each has two bends near its extremities — that at the top to form a hook, that at the bottom to hold a cord, or wire, to which a heald shaft is connected. The hooks rest upon a perforated board or a metal rack, known as the bottom board, and each is kept vertical by a horizontal needle r/, furnished with two eyes /, formed by slotting and drilling the metal ; slot e is slightly broader than a hook, and the latter is passed through it. Two plates ^, \ are per- forated to receive the forward and rear ends of and hold them in a horizontal line. Through eye / the lower end of a piece of steel wire i is passed, and its upper end is secured to a flat bar^, extending from end to end of the machine ; i acts as a spring, and thrusts the end of d through the needle plate g. A square prism k is supported by two horizontal rods /, fixed to opposite ends of the framing by bearings m, in which I slide freely. Both rods carry two fixed brackets 71, 0, the former being curved forks, and the latter bearings to receive the gudgeons of k. The prism, or cylinder ^, is perforated on each face, so that the holes are brought exactly opposite the needle points, and are pegged to hold G 82 MECHANISM OF WE A VI NG PART the cards in position ; it requires a horizontal and a rotary motion. The horizontal motion comes from two vertical rods that move freely in bearings ([ ; they carry a griffe bar r at right angles to but inclined to hooks or better Fig. 55. still, slightly inclined to both ; also bowls 5, which work between the prongs of fork n. Above r a cross-head t is keyed upon and attached by link and pins to a lever, having its fulcrum in a bracket elevated above the framing. From this lever a rod descends to a crank on the main driving shaft of the loom ; hence, as the crank V DOB BY SHEDDING 83 rotates, the rod receives a vertical movement, and imparts a lateral reciprocating motion to and also to cylinder /.'. The rotary motion is given by a catch ?/, that rests upon the top face of the cylinder, with its hook projecting beyond the forward edge of a lantern ,i, secured to one end of A;, and consisting of four rounded metal projections in line with the four corners of the cylinder, so that when ^' moves out, catch y takes hold of a projection on and pulls it through one-fourth of a revolution. It is prevented from turning too far, and assisted to turn its proper dis- tance, by two T-shaped hammers 1, each having a shank long enough to pass entirely through the dobby. Both shanks are partly round, partly square in section, and have spirals loosely threaded upon them to press against a cross- bar that connects sliding rods /, and also upon the square parts of the shanks ; hence the heads of 1 continually press upon the inner face of cylinder Ic. A chain of perforated paper cards is passed round cylinder h and held in position by two conical pegs in each face, which pass through holes in the cards. The cards are punched to correspond with the pattern to be woven, each hole equals a rising shaft, and each blank a sinking shaft. The rotary motion of h brings the cards forward suc- cessively, and the horizontal motion allows the cylinder to turn, and also moves it into contact with the projecting points of needles d. The action is as follows : — When the driving crank is on its top centre, griffe r is about \' below the upper bend of hooks &, and its top edge touches the vertical line of hooks ; bowls s are at the bottom of forks n; "cylinder k is pressing a card against plate ^, and consequently the projecting points of needles d are pushed back when blank places in 84 MFXHANISM OF IVEA VING PART the (3ards come opposite them, for springs % give way. In doing which some hooks h are pressed out of the per- pendicular by the needles, but a perforation in the card allows a needle point to enter without producing any effect on either needle or hook. The rotation of the crank causes r to ascend and take with it all vertical hooks, but all inclined hooks are left. Cylinder h moves out, when r is above the level of hooks left down, far enough for catch y to turn it without bending the needles, a fresh card is presented, and the process repeated. This dobby is neither more nor less than a small Jacquard, with one row of hooks and needles. It is only adapted for slow-running looms, for reasons given on p. 19. It is also defective at two additional points: first, because a horizontal grifFe, with a vertical lift, cannot place all warp threads of the top shed level, and leave all those of the bottom shed on the race board, but lifts all equally ; therefore, as the rear shafts must be elevated sufficiently to allow a shuttle to pass below the warp, the forward shafts move too high, and unnecessary strain results. To avoid which each shaft should have a different lift, the front one to move through the least space, and the back one through the greatest. A simple alteration in the griffe shown in Fig. 56 has rendered the machine perfect in this respect : a is the grif!e bar which swings freely upon a fulcrum pin h carried over the loom front by a bracket on the dobby. At c a hori- zontal shaft vibrates in bearings, and has two arms, e, the former connected to a by link /, and the latter to the driving crank by rod g. It is obvious that the hook nearest the fulcrum pin will move through less space than that nearest to link /, The second point relates to the horizontal traverse of the cylinder, which must be regulated to hold the needles back until the grifFe is above the hooks left down ; this is 86 MECHANISM OF WE A VING PART m m. done by making the bottom of the space between the prongs of forks n (Fig. 55) almost vertical. Careful consideration Avill show that when grifFe r reaches the same point in its descent, cylinder li again presses back some of the needles, and if any of them actuate hooks already on the griffe, those hooks will be pushed off and fall the re- maining distance; the tendency is to puncture cards, bend both needles and hooks, and generally to increase wear and tear. Of the parts added to remedy the above-named defect, the most ^ effective is separate driving for the cylinder — namely, a tappet fixed to the crank shaft, and shaped to give the proper movement, acts through levers and rods upon the cylinder, but the cost of applying it and the increased number of parts have prevented its general applica- tion. A fairly satisfactory arrange- ment is shown in Fig. 57, where hooks a are turned round, and in their normal vertical position are out of reach of grifie h ; they are pushed upon h by blanks in the cards acting on short needles each furnished with a disc instead of an eye. Another contrivance is shown at Fig. 58, in which Fk;. V DOBBY SHEDDING 87 needle a has three eyes, the first large enough to receive spring the second one considerably exceeds the dimen- FiG. 58. sions of hook c and takes its long leg, whilst the third is only large enough to take its short leg. Assuming hook c is on the grifFe and a change is necessary, a card will press back the point of needle a without pushing c off the grifFe, owing to the size of eye 2, 88 MECHANISM OF WEAVING PART but the space between the long and short legs of c is in- creased, and the natural springy nature of the wire carries the long leg off the grifte immediately the latter reaches its lowest point. Cards containing two or three rows of holes are often met with on dobbies having only one row of needles; this enables a manufacturer to produce long, striped, or bordered fabrics from a few cards, each row being cut to weave a different pattern. The cylinder is, of course, drilled to correspond with the number of rows in the card, and the dobby is said to be double-decked for two, and three-decked for three rows. There are two systems in use for bringing the different rows of holes in action — one consists in elevating the needle plate, and the other in elevating the cylinder. The latter is by far the best, for needles never work so well as when at right angles to hooks, and it is obvious that elevating or depressing the forward ends of needles and leaving their rear ends stationary, will cause one set of wires to be inclined to the other. In Fig. 55 a plan for lifting a needle plate is shown : a is a shaft that extends the length of the dobby, on which are two arms but the drawing only shows one, supporting- bars c and needle plate cj. A handle e! and a notched plate /' are employed to move g up or down. Assuming a card to have three rows of holes, the needles are opposite the middle row when handle e is in the centre notch of /. If moved to the right notch the top row will be used, and if moved to the left notch the bottom row will be acting upon the needles. In Fig. 59 it will be noticed that similar parts to those used in the last figure are attached to the cylinder instead of the needle plate : a is the shaft, h an arm, c the cylinder DOB BY SHEDDING 89 bar, d the cylinder, e the setting handle, and / the retaining plate. The cylinder motion of a dobby has been spoken of as horizontal, but some cylinders merely swing, whilst others only rotate. The parts of a swinging cylinder are given in Fig. 59. A shaft, a!, is placed near the foot of the machine and carries 90 MECHANISM OF WE A VING part v two uprights c, into the upper ends of which the cylinder d and its bearings are secured. An arm c/', with a long curved slot, swings freely upon pin d in the lower part of the machine framing as bowl fitted upon lifting lever ^, moves up and down the slot. A connecting rod li couples and c and conveys the motion of the former to the latter. Other things being equal, a swinging motion is not quite so good as a horizontal one, because it moves in the arc of a circle and causes the upper parts of the holes to come opposite the needles when contact is first made between card and needles ; but as the cylinder arm moves into a vertical position the needle points occupy the centre of the holes, whereas a horizontal motion constantly holds the hole centres opposite the needle points. A rotary cylinder motion invariably requires lags instead of cards, and a series of springs a (Fig. 60) (in the drawing given they resemble tuning-forks), the handle of each is screwed to rail ft, and one prong rests against needle c, the other against lag d. When cylinder e is turned, a peg presses the prongs of a together, and hook / is moved in range of griffe g. The forks when in position form an almost continuous line across the machine and give comparatively large surfaces for the pegs to act on ; this is essential, as it would be next to impossible to ensure perfect working without them, for a peg would slip over, under, or at one side of its needle. A modification of the last method is shown in Fig. 61. Flat springs ^ wide are screwed upon rail h and slightly bent outward by rod c. Needle d is riveted to spring a, so a peg in 6 presses back a and d and forces hook / over griffe g. Both methods are moderately satisfactory, but of the two the last described is preferable. Fig. go. 92 MECHANISM OF WE A VING PART As a fresh set of cards is required for each pattern, they form the most expensive item in connection with dobbies, and it is not to be wondered at that machinists and Fig. 61. manufacturers have devoted considerable time to the problem of reducing the cost of using them. A continuous band of perforated paper, also of wire gauze, and canvas cloth, having some of the perforations stopped V DOB BY SHEDDING 93 up with varnish, or other material capable of resisting the tendency of needles to break off, or puncture it, have been tried with little or no success ; but lags, consisting of pieces of wood (each drilled so that its holes coincide with the needles), are formed into a chain by driving staples into their sides near each end, and linking all together by rings ; then wooden pegs are pressed into some of the holes to serve as risers or sinkers, and lags furnish manufacturers 12 3 4 O [( Fio. 62. with a successful rival to cards, as pegs can be pulled out and readjusted for a new pattern. Lags are altered in form to suit the make of dobby they are intended to be used with. Some have single, whilst others have a double row of holes. Where the latter occur, holes forming the second row are exactly midway between those of the first row, hence the pegs when inserted have a zigzag appearance and allow each lag to govern two sheds instead of one, thereby reducing the length of lattice required for a given pattern. Pegs for all double and many single rowed lags are cylindrical in form, 94 MECHANISM OF WE A VING PART and usually of wood, about |'' long by |" in diameter, but as such thin wood is liable to break off when working, metal has been tried as a substitute with, however, only partial success (see Fig. 62, No. 1). Other pegs for single pick lags have a round shank and an oblong head ; needles are then dispensed with by causing the pegs to act direct upon the hooks so as to push them on or off the griffe ; in the former case a peg lifts a shaft, in the latter it depresses 2 oo«o«»»o/oo |000»0©»«/0 0 O o/ •000 «0«»8 O O /• •O0O«O« 7 * •o»«oooo/oo; 30«0«*«00/0 0 o o/ •••ooo«oS oo/o»»«ooo» 5 4- 2 7 5 3 X 0 - ^ 0 _ - ^ ^ 800 0 " ^ 0^ _ - ^ ^ '800 0 0 799 0 ^ 0— " 0 ^^-^ 0 400 0 ^ - - " o-- — 2 S ^ - ^ ^ 399 0- 0^" " Fig. 111. Split or Double Scale Harness A harness known as the "Bannister" is in extensive use for weaving wide patterns in a fine reed, from a small Jacquard. It has a comber board (Fig. 112) fixed about 12'' above the warp, and couplings with top loops 16'' long are drawn into it, but all are previously knotted 8" above the mails at a. A strip of wood or metal, thick by 1|" deep, with rounded edges, is pushed between the loops of all couplings in one row of the board and just below the knots. As a consequence, the number of shafts employed 204 MECHANISM OF WE A VING part corresponds with the number of holes in one row of the Fig. 112. board, viz. from 8 to 24. Two, three, or even four VII THE FIGURING HARNESS 205 consecutive couplings in each repeat are moved by one hook, but they can be lifted separately by the shafts in any desired order in the following manner : — Let it be assumed that 16 holes form one short row, then couplings 1, 2 are tied, or w^arped, upon the neck cord of number 1 hook ; 3, 4 to number 2 hook, etc., or as occasionally happens, couplings 1, 3, 2, 4 are respectively tied to hooks 1, 2 ; this is done to enable a Jacquard to lift plain cloth sheds without the aid of shafts ; but as couplings 1, 2, 3, 4 form part of different longitudinal rows, they are threaded upon four separate shafts which may each be moved in any desired order by spare hooks in the machine. If cards fall over warp, hooks should be provided near both ends of the cylinder, so that cords from them may pass through the comber board and down to opposite ends of the shafts without crossing. This arrangement also adds equal w^eight to each end of the grifFe. Cards are only perforated where figure is to be formed, therefore a shaft, or shafts, must lift some warp left down by the Jacquard, for the purpose of interlacing it with weft in the ground portions of the fabric. One card is required for each pick. The advantages of this harness are : first, a pattern woven from a set of small cards can be doubled, trebled, or quadrupled in width ; second, difi'erent ground workings can be woven with the same figure if the shafts are moved by spare hooks placed near enough to the cylinder ends to allow short cards to work outside the figuring ones ; third, less cost incurred in designing, because only figure requires attention. Its one disadvantage consists in giving a figure a slightly rough, defective outline by moving warp threads in pairs. From 8 to 24 shafts are often manipulated by attaching 206 MECHANISM OF WE A VING PART two neck cords to every hook, and passing them through contiguous rows of holes in a bottom board fixed 2'' to ?>' below the hooks ; it contains two rows of holes for one row of hooks in the Jacquard. Each neck cord is knotted at two points, about ^" apart, to provide eyes for the reception of short shafts ; the latter are then out of the way ; they work in the least crowded parts of the harness, and give satisfactory results. Another double scale harness has been in common use in the cotton industry for many years. It is built to effect a saving of half the cards as compared with an ordinary harness. Two separate machines are often used — one takes all odd-numbered couplings, the other all even numbers ; but a double lift single cylinder, or a double lift double cylinder machine answers the purpose admirably if the neck cords are disconnected, and tied to separate couplings in the usual manner ; then if cards are presented to needles, with holes cut in all possible places, plain cloth will be woven, for all hooks on one griffe move alternate couplings, therefore the first card on a 400 machine will lift all odd threads from 1 to 799 in every repeat, and card 2 all even threads from 2 to 800. Any kind of figure in which only one set of threads is to float can be obtained ; and all grounds in which no two adjoining threads are required to lift simultaneously are also possible. It is obvious that a warp float must be made by weaving a fabric face down, and a weft float by weaving it face up. Pressure Harness A pressure harness requires Jacquard and healds com- bined, but both to be moved independently, and yet to act upon the same threads. A figuring harness may be VII THE FIGURING HARNESS 207 straight, centred, combined, or of any other type, but it must be placed farther from the fell of cloth than for common work. As a rule, a space is left between back shaft and comber board of T for cotton, and 10" for linen warps; from two to five threads are drawn through each mail, but where more than two are needed, decked mails must be used to prevent the threads from twisting. Heald shafts, corresponding in number with the threads in one repeat of the ground pattern, are suspended as near the fell of cloth as possible, and are either furnished with large eyes from 2^'' to 3'' deep, or double the number of clasped healds are used ; in the latter event, every thread is drawn above one clasp and under another. Warp which has passed in groups through the mails is divided, and drawn singly in straight order through the healds. When all eyed shafts are level the Jacquard must be capable of lifting warp without obstruction from the healds, after which one shaft must sink and one rise for every pick in order to weave a satin or twill. A sinking shaft pulls down one thread from each repeat of satin lifted by the Jacquard, and a rising shaft lifts one thread of each repeat left down by it. If the rising and falling of these shafts is suitably arranged weft satin will be formed throughout for ground, and warp satin for figure. The saving efi'ected by this harness is very great; for example, a 600 machine with 4 threads to a mail and 4 picks to a card gives a pattern that repeats on 600x4 = 2400 threads. This number moved by a single harness would require four 600 machines, and therefore 4 cards to a pick, in place of 1 card to 4 picks, or altogether 4x4 = 16 times as many cards as a pressure harness. The smallest possible sheds and shuttles are employed ; 208 MECHANISM OF WE A VING PART nevertheless, it puts great strain upon warp ; but, on the other hand, it is firmer and better adapted for very fine fabrics than the Bessbrook, or any other twilling machine that has been introduced to replace it. All figures produced Fig. 113. by it have a somewhat jagged appearance, and lack fineness of outline and detail. Healds are moved by the Jacquard, by tappets, and by dobbies. When the former is employed, each shaft is suspended by tying one neck cord to two hooks and passing it under a grooved pulley a (Fig. 113), from which cords VII THE FIGURING HARNESS 209 go down to an upper shaft. A shaft can be placed in three positions : when both hooks are down ; when one hook is lifted ; c?, when both are lifted ; the first holds warp down that would otherwise be lifted by the harness ; the second puts a shaft into its out-of-action position ; and the third lifts warp that the harness would leave down. But this system only lends itself to the production of unequal fabrics, in which warp exceeds weft in the proportions of 2, 3, 4, or 5 to 1. If healds are moved in a similar manner by tappets or dobbies, the Jacquard can remain lifted for any number of picks, and shafts may move pick by pick to weave the satin. Gauze Weaving Designing and fabric structure are outside the limits of the present treatise ; nevertheless, it may not be amiss to briefly describe the nature of the work to be done in order to produce fabrics in which any portion of the warp threads is made to twist partly, or wholly, round other threads. Nearly all classes of shedding motions are available for this purpose, but some answer better than others. In every case special appliances are required to give the twist, such as healds, a reed, or needle frames. The former are in most extensive use for all round work, but the two last named possess advantages for weaving those fabrics where the order of lifting and twisting is limited. Every twisting thread must be moved at two points. If healds are employed throughout, ordinary ones are used to lift a straight or open shed, whilst a heald and half heald com- bined cause one thread to twist half round another and form a cross shed. The accompanying Figs. (114, 115, and 116) show how this effect is produced. Fig. 114 is a draft and tie-up for % P 210 MECHANISM OF WE A VI NG PART simple gauze ; the thick, short lines at the top indicate that two threads pass through one dent of the reed. Horizontal lines 1, 2 are two heald shafts ; the former containing all even threads, the latter all odd threads of warp ; 5, d are respectively standard and doup ; both are better seen in Figs. 115 and 116. Standard s is an ordinary heald shaft into which the half heald d has been looped by passing the twine over and through the eyes of s. When warp has been drawn into shafts 1 , 2, alternately, every odd thread I 1 i \ 1 \ \ \ V \ s 5 C 0 Fig. 114. is crossed under the next even one, and passed through a loop of the half heald. Dots on vertical lines o, c (Fig. 114) represent the order of lifting the shafts ; o is an open shed obtained by lifting 6?, 2, as clearly seen in elevation (Fig. 115) ; c is a cross shed made by raising c?, 5, also shown in Fig. 116. Twisting one thread round another and lifting the same threads for every pick are distinguishing features of gauze ; the twist binds all threads on shaft 1 firmly into the fabric, although they are never lifted above weft. It is quite evident that threads lifted for a cross shed will be strained VII THE FIGURING HARNESS 211 by passing crossing threads over a bar a (Fig. 115), that moves in to slacken twisting warp, and stationary threads over a fixed bar h. s Fig. 116. When two ends cross two, or more, a method is some- times adopted which dispenses with a standard and renders 212 MECHANISM OF WE A VING PART unnecessary the use of a back shaft for crossing threads. Two half healds c, d (Fig. 117) are employed with their laths at the top ; a loop from each is connected by a glass bead 6, and threads 1, 4 pass underneath 2, 3, and through bead e. Threads 2, 3 are drawn singly through an eye on a or 5, then all four are put into one dent of the reed. By lifting c, d in rotation, c causes the bead e to pull threads 1, 4 to the left of 2, 3, and d acts to pull them to the right of 2, 3. Tie 5, 6 show the above, and 7, 8 show 12 3 4 \ 1 1 : i f — \ ( E 1 5 6 7 8 Fig. nr. threads 2, 3 working in plain cloth order, with 1, 4 twisting round them as before. Coarse open fabrics can be woven without healds by the application of two shafts a, h (Fig. 118), into which needles c, d are driven ; the needles in each shaft must be close enough to present an eye opposite every dent in the reed. Shaft a is placed above the warp with eyes c at the bottom ; h is placed below, with eyes d uppermost, and a thread of warp is drawn through each. When fixed in the loom so that all warp forms one line, the eyes of shaft a VII THE FIGURING HARNESS 213 are midway between those of shaft h ; hence, by forcing a down and h up warp will form two lines to receive the shuttle; they then move back again until eyes c, d have separated warp about f " in the opposite direction. At this point lateral movement is given to both ; a slides in one direction, h in the other, but only just far enough to cross D I B Fk;. 118. threads. The next vertical movement again forms a shed, therefore, if the first was an open one, the second will be crossed. Lateral motion in both staves is derived from cams upon the main driving shaft. One, situated near the end, is in the form of a disc, and the other is a cross- grooved cam secured near the centre. A better plan is to employ a gauze reed that is made to rise and fall by the action of a cam. It has a half dent 214 MECHANISM OF WE A VING I'ART placed between all ordinary ones ; the half dent is provided with an eye near the top for a crossing thread to pass through, but straight threads are drawn through eyes in an ordinary heald, or through mails in a J acquard harness, and are then taken between full dents of the gauze reed, and c finally, straight and crossing threads go in pairs between the dents of an ordinary reed. When all straight warp is above the half dents, the back harness is pushed or pulled sideways to carry its warp to the opposite side of the half dents, the reed then rises and takes up all crossing threads on the right or left of station- ary threads. VII THE FIGURING HARNESS 215 Gauze Harnesses Special harnesses are built to weave gauze fabrics with bottom and top doups ; still gauze can be woven with an ordinar}^ harness if doup and standard shafts are placed in front of it, and if crossing threads pass over an easing bar that moves to slacken them, whenever doup and standard rise together. Movement may be given to an easer by spare hooks from the Jacqaard, or by a tappet fitted upon any convenient part of the loom. An illustration of each method is given in Figs. 119 and 120. Fig. 119 shows a lever a, to which a cord is tied and carried to the Jacquard ; a is fulcrumed at &, and at its extremity a rod c passes across the loom to support all crossing threads. An upward movement of the cord carries c down and slackens the threads. In Fig. 120 the cord goes down to a tappet treadle, and rod c is below the straight threads, hence easing is accomplished by pulling the cord to elevate c. Every crossing thread for a bottom doup, with two ends in a dent, after passing through a figuring mail, is taken under its fellow, and also drawn through a separate loop B Fig. 120. 2l6 MECHANISM OF WE A VING PART in the doup, then, for alternate picks, standard and doup go up together to lift half the warp, say all odd threads. At the same time some even threads are lifted by hooks to form figure, because wherever both odd and even threads are lifted in any dent, no douping can take place, for the simple reason that there is nothing to cross round ; but where even threads are down, odd ones are crossed. On intervening picks, lift the half heald, together with crossing threads by the figuring harness where gauze is wanted, and odd and even together to form figure, or even only for plain cloth. A harness of this kind can be used to weave many gauze fabrics ; but it is limited in the sense that only one pick can be inserted into a shed in the gauze portion, if at any other part plain cloth is used. Bottom Doup Harness If two or more picks in a gauze shed are combined in any way with plain cloth a special gauze harness is required. These are built in various ways ; generally there are three distinct sections, and three separate comber boards. If cards fall over one side of a loom from a 600 machine, and the reed has four threads through each dent, two cross- ing two, the hooks must be divided into six parts, and one-sixth used for easers, four-sixths for figuring, and one- sixth for douping, or, 8 rows and 4 hooks from the back for easers, 33 rows and 4 hooks for figure, and 8 rows and 4 hooks for doups = 100 + 400 -f 100 = 600. In the event of cards falling over warp take the two front rows for easers, eight middle rows for figure, and two back rows for doups = 50x2 = 100, and 50x8 = 400 = 100 + 400 + 100 = 600. In building an easing harness neck cords are attached VII THE FIGURING HARNESS 217 alternately to levers in two lines a, h (Fig. 121), one □ Fic. 121. approximately 4|'' above the other ; and mounting 2l8 MECHANISM OF WE A VING PART threads from the levers are connected to couplings c, the mails of which are in a lower plane than those of the figuring harness, l)y from 1" to 2V', but on an average \y'. Instead of the arrangement described above, Devoge & Co. employ two griffes for a single lift machine ; but the second, which is capable of adjustment, has a shorter traverse than the first, and merely controls easing threads. These makers also move doup and easing hooks by one set of needles, for the purpose of ensuring certainty in lifting both together and slightly reducing the cost of cards. Metal mails are often used, still glass ones are preferable, as they do not cut, and are smoother ; they contain all crossing threads, and lingoes must be heavy enough to pull them down, viz. from 4 to 10 per lb. Any one, or any group of the 100 in each repeat, can be lifted separately to slacken the threads each controls. Two stout rods 6, called the bridge, touch the under side of all straight threads, and are bolted on slotted brackets that permit of lateral adjustment. As a rule, they are from 3'' to ^' apart. The figuring harness / is about \T in advance of the former and is built on the ordinary plan, but fine enough to receive every thread of warp separately. Between doup harness cj and figuring harness / a space of from ^ left to minimise the strain when threads cross. The mail eyes of this section are about lower than those of /, and sometimes contain two eyes for doup twine to be drawn through, as shown at h. At other times a single eye i is used and doup twine is pushed through it ; this, however, is troublesome to the weaver when warp breaks, as doups are apt to fall down out of the mails. On the other hand, if new doups are required that pass through two mail eyes, it is necessary to build them at the loom. Doup slips are all VII THE FIGURING HARNESS 219 made fast to a heald shaft which is raised and lowered for every pick by making it fast to the machine griff e and pulling it down by springs ; it requires careful adjustment, or the doups rapidly wear out. Wearing is to a large extent prevented by fastening regulating cords upon opposite ends of the doup shaft, and also to the comber board, and so prevent the springs from pulling leashes tight upon doup mails. Doup mails vary from 16 to 24 to the lb. After crossing warp is drawn double through easing mails c, and all threads singly through figuring harness /, they are separated, and the first pair of crossing threads are passed to the right or left, but under the first pair of straight threads, then drawn through number 1 doup loop, and finally, straight and crossed, go through the reed 4 in a dent. It will be noticed that crossing threads are drawn into easing, figuring, and doup mountings (Fig. 122), whilst straight threads only go through one eye in the figuring portion. To weave figure lift half heald 7 and figuring threads / in the middle harness, as at 9. To weave gauze lift those easing mails c that slacken the threads required to cross, corresponding doup mails ^, and half heald j for cross sheds, as at 10. For open sheds lift half heald 7, and cross- ing threads in middle harness/, as at 11. With such a harness it is next to impossible to reach the limit to which variations can be carried ; the crossings may be of a most irregular description, including one crossing one, and six crossing six. Picks in one gauze shed can be altered at pleasure. Any single doup, or any assortment of doups, can be brought into use at any time. Warp and weft floats in the figure pattern can be combined with all kinds of ground working, provided plain cloth surrounds 220 MECHANISM OF WE A VI NG PART the gauze, for it is essential that warp shall be thoroughly opened out before figuring is attempted. It has been elsewhere mentioned that double-acting shedding motions — namely, those that give a semi -open shed — are not so well adapted for gauze weaving as closed ^ % 3^ 5 6' r 1 ( ( -^^ -< H ) 5 \ — -< >- — ^ '. . . -< ) H ) -^^ -< )- 5 -< ) JS \\ — 56e-^ < k 0 ^ 11 ) ) Fig. VA2. shedding motions. Still any type of double lift machine can be used if suitable mechanism is applied for lifting standing threads half-way, to enable crossing threads to twist round them. It would appear that, as one shaft begins to rise immediately another begins to fall, slack warp given off by a sinking shaft would be taken up by a rising one ; but it must be remembered that both shafts are level midway vir THE FIGURING HARNESS 221 between the top and bottom lines, and the easer, which was in its most advanced position when the downward move- ment began, has moved back at least half-way, and taken up a considerable length of the crossing threads, therefore such threads, at this point, must either pull up standing ones or cause breakages. It is the function of shakers to prevent strain and break- age by lifting standing threads half-way, and dropping them again during the formation of each gauze shed. For heald work a shaft a (Fig. 123) is held parallel with the heald shafts by brackets, and carries three arms, two of which h are connected by cords to those shafts only that contain standing threads ; the third arm c has a rod attachment d to one of the con- necting arms of the driving crank. As the latter rotates, a vibrating motion is conveyed by rod d to shaft a and through arms & to the standing shafts, which is sufficient to slacken the warp and permit of Fig. 123. It is equally easy to apply parts to a Jacquard harness that will give similar results. Probably the best shaker consists of Bannister shafts threaded through the couplings which contain the standing warp, and elevating them every pick by means of a tappet. 222 MECHANISM OF WEA VING Top Doups Top doups are extensively used in the cotton industry for shaft work as well as with Jacquards. A top doup harness is similar to one for bottom doups, except that the half heald shaft is uppermost, that doup i 2 34567 8 — > \ )- ( K ) J e— i K ) (— f — ^ -< ) ) i )- -< i k y- y -i ? ? — \ \ — 1 ' 1 [>< A Fig. 124. lingoes are somewhat heavier, and that crossing threads are above, instead of below, straight ones. The half heald shaft is sometimes screwed to the doup comber board; at other times it is suspended from the machine. In the former case lines 1 to 8, in Fig. 124, are intended to represent threads passing through mails, the couplings of which are in the first row of holes in comber board h/ VII THE FIGURING HARNESS 223 3, 4, 7, 8 are crossing threads, but they also pass through mails in easing mounting a, and through two loops of half heald that form part of doup mounting c. An open shed 0 is formed by lifting doup mails with straight threads in harness h wherever gauze is required. For a cross shed d lift easers with straight and crossing threads where necessary ; and for figure / lift doup mails with figuring threads in harness h. The chief differences between the two systems of work- ing consist in lifting crossing threads for every pick of gauze with a bottom doup, and straight threads every pick with a top doup ; also with the former, doup and easer in- variably move together ; in the latter, this never takes place, for lifting a doup mail allows crossing threads to assume the parallel position, whilst the weight of doup mails is sufficient to hold them in the cross, but lifting an easer takes off the strain and permits straight threads to rise at the opposite side. A top doup has many advantages over a bottom doup, chief of which is that doups are in the most convenient position to repair in case of breakage or disarrangement ; as it is impossible, even with single-eyed doup mails, for doup slips to fall out of reach of the weaver. Top doups may, of course, be used when twisting threads go under straight ones, if the former are lifted every pick ; but in such cases doup twine and stationary warp rub in the formation of an open shed, the doups also have a tendency to lift straight threads above the bottom shed line and form irregular openings. Doup Harness in Two Sections Another form of doup harness is still occasionally met with ; it only requires two sections in the Jacquard — one for 224 MECHANISM OF WEAVING part doups and easers^ the other for figure. Every doup slip a (Fig. 125) is tied to a separate lingoe &, and a second thread c from the lingoe goes up through comber board and is VII THE FIGURING HARNESS 225 tied upon figuring thread e, that controls the same warp as the doup slip. By this means a slip is only lifted when its warp is lifted in the figuring harness. Doup mounting / is of the usual type, but shortly below the neck cords a mounting thread \ \ 5 ^ > < D ' ' ) ir- e- ) — ) i — -X < ) — > 5 — ^ ) — ( € y- y- — y — c > D Fig. 126. g is led off to the easing comber board li. Although this plan augments the figuring capacity of an ordinary Jacquard, it gives the harness an exceedingly crossed appearance, and increases friction. Fig. 126 shows two threads crossing two ; threads 1 to 8 fill part of a row of comber board d ; 3, 4, 7, 8 are crossing threads that pass through two easers and two doups ; 0 shows the lifting for an open shed, and c that for a cross shed. Q 226 MECHANISM OF WE A VING PART Wilkinson's Harness Wilkinson patented a harness that performs the work 2 2 ^ G Fig. 127. of an open shed Jacquard without requiring additional VIII CARD-CUTTING 227 lift in the griffe. It is attached to a double-acting single cylinder machine, in which one neck cord a (Fig. 127) is connected to two hooks ^, c, and passes under a grooved pulley mounted in a pair of plates, one of which is shown at e. The plates also support a second grooved pulley /, round which a supplementary neck cord g is passed and secured to one of a series of fixed bars ft, running lengthwise of the machine and between alternate lines of hooks ; the other end of (j is made fast to the mounting threads. If both hooks are down, warp governed by them is on the bottom shed, but by lifting one it is raised to the top line, and may be retained at that point for any length of time, by causing the reciprocating griffe bars to carry up one hook at each upward journey, for cord given off by the fall of h is taken up by the rise of c without putting addi- tional strain upon any working part of the machine. PART VIII CARD-CUTTING Designs are painted on paper ruled vertically and hori- zontally so that every space between two vertical lines shall represent a warp thread, and every space between two horizontal lines a pick of weft. Paint put upon the small squares, formed by vertical lines intersecting horizontal ones, denotes, as a rule, warp lifted above weft, and therefore holes in cards. At certain regular intervals thicker up and cross lines, called bars, are ruled ; the up lines enclose as many threads as there are needles in one short row of the Jacquard to be used. Cross lines mainly serve as guides to the designer and card-cutter. 228 MECHANISM OF WEAVING PART Each card is cut from one horizontal reading of the design extending from side to side, but holes punched in cards have the effect of turning up the design, bar by bar, into short vertical lines. Card-cutting machines are of two classes — namely, those used to transfer a pattern from design paper to cards, and those used to copy perforations from a set of existing cards to a new set. In many cases they are distinct, but occa- sionally they are combined. The most familiar machines of both classes appear to be of French origin. The oldest of all is still used in districts where small Jacquards are the rule. It consists of a pair of perforated plates that are hinged together, and after placing a blank card upon the lower plate, the top one closes over it, and is locked by a sliding catch. Both plates are secured to a frame, with rollers underneath, that runs on railway lines. A third, or carrying plate, is perforated like the others, but furnished with a handle at each end ; it is placed on a bench, above which the design to be copied is fixed, and a box containing round punches long by in diameter, and ^ heads, is within easy reach. The punches are dropped one by one into holes of the carrying plate to correspond with the painted design, then plate and punches (for the heads prevent the latter from falling out) are fitted above the other plates, and pressure of some kind is exerted to force all punches through the card. The apparatus now generally used for this purpose is a roller press, turned by one hand, and the plates are pushed beneath it with the other. The process is much more expeditious than might be supposed, especially where the groundwork of a pattern is regular, such as plain cloth or twill. In the case of a plain ground the punches are set for the first card, but after that VIII CARD-CUTTING 229 a cutter cuts odd picks only for the first reading, because few punches re- quire changing, probably not more than a dozen. On the second readino^ he takes all even picks. A twill ground is read by repeats of the twill ; thus for a four-end twill every fourth pick is cut for each reading. The above machine is converted into a very use- ful repeater by the addi- tion of a frame that sup- ports a carriage on sliding bearings. The carriage a (Fig. 128) generally con- tains 612 horizontal needles 5, arranged in 12 rows of 51 each. Every needle is supported at the front by two perforated plates c, the former is forced away from the latter by spiral springs threaded on four spindles, all riveted to c, but pass- ing freely through d. A similar perforated plate /, fixed in the rear of carriage a, supports the back ends 00 p n n n n n n n n n n u uu u u u u u u u u u X 230 MECHANISM OF WE A VING tart of h. Each needle has a shoulder fitted at a certain distance from its point by coiling a piece of wire once round, then a thin brass spiral / presses against the collar at one end and against plate / at the other, to constantly hold the points of h beyond plate c, A punch box g is drilled to receive all the punches, which are pushed in head first, and prevented from passing out at the back by a thin brass plate A, also drilled, but with holes only large enough to permit needles & to enter. A stud pro- trudes from cj near each end, to hold carrying plate i, with its holes opposite those of g. Then the card to be repeated is fastened by its peg-holes upon cylinder pegs in plate A, carriage a is drawn forward by pressing down a treadle, and springs ] are strong enough to cause needles /> to push punches from box g into plate % where holes are cut in the model card, but the card causes ^ to contract and needles to slide back that are opposite blanks. Carrying plate I is removed to book plates containing the card to be cut, and all are moved under a press either turned by manual power or by an engine. After cutting the card and replacing carrying plate i upon the studs of ^, a comb, consisting of 612 pieces of wire driven into a wooden back, drives all punches into punch box ^, and a second card is fixed upon the pegs of li to continue the operation. This machine is good, reliable, and simple, but compared with automatic repeaters, the process is slow and costly. Eeading-in Machine The first card-cutting machine was patented in England in 1821 by Wilson from a foreign communication. It is still used in many parts of the country, but is being gradu- ally pushed out by modern improvements. VIII CARD-CUTTING 231 by passing each through a bead and hanging a heavy lingo e c upon it. At the back of the machine cords a are led 232 MECHANISM OF WEAVING PART down in a straight line, separated by a comb d into sets of threads corresponding with the number of vertical spaces enclosed by thick lines on the design paper, and further divided by rods to form an end and end lease. At the front each cord a goes through the eye of an ordinary Jacquard needle /, the latter are arranged in rows of 12, and are supported by needle board and heel rack. Imme- diately in front of the needles a punch box and carrying plate % as previously described, are placed. It therefore follows that by drawing any cord a forward, its needle will push a punch out of box g into carrying plate i, and by means of the latter it is taken to a railway press and forced through a card. The operation is as follows : — The design is placed before the reader, a straight-edge assists him to read across a horizon- tal line, and he proceeds to transfer the pattern to cords a by weaving short cross twine or picks amongst them, in such a manner that every vertical thread representing lifted and sunk warp must respectively pass in front of and behind a cross twine. As reading continues cords a are gradually pulled onward, until those picks first read-in are carried to the front, where another man inserts a roller j in place of a pick, fixes it in a sliding frame, presses down a treadle, and draws forward all cords in front of 7, and by so doing, pushes corresponding punches into the carrying plate. Reading-in and Eepeatinct Machine In the course of a few years the foregoing machine was converted into one for reading-in and copying, from which we obtain our best automatic repeaters. Instead of using endless cords, a series of vertical cords a (Fig. 1 30) were each formed with a loop at the bottom, and after pushing a rod VIII CARD-CUTTING 233 tlirough all, the latter was forced into a slot in roller h. Near the top oi a 2, second series of loops c were made, and each strand threaded through a separate hole in guide board Fm. 130. then finally dropped into a wire hook ^, which in its turn was retained in a fixed frame / by threading its con- necting cord g through two holes in /. Cord after passing over a guide pulley, and through comber board A, terminated 234 MECHANISM OF WEA VING PART with a lingoe i. A cord j was led over two guide rollers to the vertical part of cord /.', and there tied. The coupling twine of lingoe m formed part of and was drawn separately through holes in board /, then taken up to and over a glass rod, bent at right angles, and made fast to needle n. Three perforated plates j?, ^, r served re- spectively as front and back guides for needles n and cord h A collar twisted round n acted as a stop-hoop for spiral 0, and its rear end abutted against plate ^. Punch box 5 and carrying plate t were fixed immediately in front of needles as already described in the reading-in machine. The weight of m contracted 0^ and pulled needle n back until such time as the design, read into cords had to be transferred to cards, when, by inserting a roller in the posi- tion of the top cross cord, and pulling it forward, cords g, y, k would be elevated, the weight of m removed from springs 0 would expand, and drive some of the punches into carrying plate after which it was ready for the press. This completes the reading-in section of the machine. The repeating section will be readily understood. Over the centre of comber board / an ordinary 600 Jacquard was placed, and a thread u from each hook was connected to its proper cord h By hanging the cards to be copied in a card cradle and leading them round the cylinder, the machine, when set in motion, would lift cords k corresponding with holes in the card, and in each case a spring 0 would force its punch into the carrying plate t Vertical Punch Reading-in and Repeating Machine An alteration was afterwards made in the form and position of the punches, by means of which an excellent machine was obtained. All parts of the reading-in section VIII CARD-CUTTING 235 remain precisely as in the former machine ; they are lettered in Fig. 131 from a to ']. The Jacqnard and connections n also remain unchanged ; but cord /i', after passing the glass Fig. 131. rod, is taken over a guide pulley thence down to the top of a vertical punch ?i, approximately IT' in length by in diameter. Each punch is filed away on one side down to 236 MECHANISM OF WEA VING part its diameter line to form slots jp, ^, both deep, and a piece between them of \' is left round. At o a space of is filed away at both sides, leaving a thickness of only, to receive a fixed comb with 53 teeth, all ^' deep, and thick enough to touch two punches without preventing them from moving freely ; this comb stops punches n from twisting. Slots |7, ^ in two adjoining rows face each other, so that a movable comb 5 may be pushed through slot ^, and each tooth will lock two rows of 12 punches ; s has 26 teeth, deep, and sufficiently wide to rest upon both lines of punches, and still allow it to slide in and out amongst them. After withdrawing 5, any of the cords, j or may lift some of the lingoes m, when punches, normally sup- ported by the weight of m, will fall, until slot faces the teeth of s, then as the latter moves forward some punches will be locked, with their ends protruding through plate t. A blank card is placed in position upon perforated plate ii\ the compound lever v is operated, and id is carried up to touch t On reaching the latter point every punch that has a tooth of comb s through slot ^ will puncture the card, whilst those locked at slot ^ are out of its reach. This machine is extensively used on the Continent, and it will be noticed that few alterations are required to bring it into the form in which Devoge & Co. supply it. Devoge k Co.'s Automatic Eepeater This machine is a repeater only, therefore all the reading-in parts of the former are cut away and the Jac- quard and its connections alone left. A 600 Jacquard is fixed on an iron framing above comber board v (Fig. 132), and cords % go down to lingoes m. Cord k is attached to u ; it passes over a roller /, and is made fast to punch n ; the VIII CARD--CUTTING latter only differs from the French punch in having the metal filed away on both sides of slots jp^ c[ (Fig. 131), instead of leaving the punches half round at ^, ^. By adopting this form of punch the makers believe increased steadiness will be given, because comb s supports each on Fig. 132. both sides. Cards have peg and lace holes punched by a separate machine, after which they are laced into a contin- uous chain, passed over guide cylinders and between plates ^, w. The Jacquard rises, comb s, which has as many teeth as comb r, is withdrawn, punches n fall where their lingoes m have been lifted ; comb s, in moving forward, passes through 238 MECHANISM OF WEAVING TART slots c[ of lifted punches, and slots ]) of sunk punches, to lock them before cutting plate %\) rises with the card to be cut. Plate n:> is moved up and down by eccentrics, and the Jacquard by a positive cam ; these render the machine thoroughly automatic. M'MuRDo's Eepeater M'Murdo's machine is the most recent modification of the French type. The maker has succeeded in doing for a repeater what the Jacquard does as a shedding motion — namely, giving all the parts a direct action, and thus obtaining a more compact machine from fewer parts. A 600 Jacquard is fixed over the punch box, in which cranked needles long govern 18'' hooks ; the latter are turned to face the spring box, and grifie bars are set to miss all vertical hooks and take up all inclined ones. Two pieces of wire %i (Fig. 133), when combined are 1Y' long ; they connect hooks and punches, n is flattened at the bottom, punched, and bent at right angles, wire u is similarly treated, then a steel spring 2 is pushed over it ; n is threaded through the punched hole of and %i goes through that of ii^ with spring 2 between the two bends, thus forming a sliding joint. Punches n are IT' long, and notched as usual at three places, but only on one side; the top notch o is V^' long, the two lower ones are each J'', whilst the round portion between them is \' long. All slots in two adjoining rows face each other. Combs r, s both have the same number of teeth — namely, 26, the former serves as a stay, also to prevent twisting, but 6- is the movable comb. Blanks in the cards to be copied, in pushing back needles, press hooks over the grifFe bars, comb 8 is with- VIII CARD-CUTTING 239 drawn, punches corresponding with hooks on the grifFe are lifted out of reach of the blank card on cutting plate u\ after which comb s moves in to lock them either up or down, and the punch box is brought down bodily, by means of eccentrics, upon the stationary plate \l\ Sliding joints 2 permit of the downward movement. This machine is also per- fectly automatic and reliable. Nuttall's Eepeater About twenty years ago Nut- tall introduced what many be- lieved to be a good automatic repeater, but it has not proved > a striking success, although its construction is both original and ingenious. The upper part of this machine contains needles a (Fig. 134) re- sembling those in a Jacquard, the cards to be copied are passed over a cylinder d, and moved successively into contact with them ; a hole in a card leaves a needle a unmoved, but a blank pushes it back. A set of vertical wires &, of which 51 I W \ 1 w Fig. 133 240 MECHANISM OF WE A VI NG PART move on each fulcrum pin are coiled at their centres to connect and control another set of needles c, by giving them an opposite movement to that imparted to the first set. Each needle c is furnished with a cylindrical plug; that in its normal position covers the head of a punch /, and prevents it rising. Every punch so covered will be forced through the card to be cut at the next upward movement of cutting plate \ but a blank in a model card withdraws the plug and allows punch / to rise ; its weight alone resting on the card not being sufficient to make a perforation. There are 12 horizontal rows of plugs, each row contain- ing 51 ; they are placed in tiers along a stepped, perforated plate \ so that the bottom or short row of needles c governing them are connected to the top row of a, whilst the top row of c, viz. the longest, are coupled with the bottom row of a. Beneath the plugs are 12 rows of VIII CARD-CUTTING 241 punches, with heads varying in length to suit the posi- tion of the plugs which act on them. All are provided with a collar i, formed at one uniform distance from the cutting point, and they are supported upon plate 7 as )i moves down to bring the next blank card into position. The machine failed partly through the liability of its parts to vibrate, partly through the superabundance of wire employed in its construction. The thickness of a plug made all the dilference between forcing a punch clean through a card and leaving the latter blank ; in some cases vibration prevented a card from being perforated ; in others, the setting was so close that cards were partially cut where they should be blank. The so-called Eoyle repeater is in all essentials the same as Nuttall's machine. At least three attempts are at present being made to produce a machine that will read cards from the design automatically, but up to the time of writing, none of them can be said to have passed the experimental stage, notwith- standing the fact that one Company has been in existence for the last eight or ten years for the purpose of supplying the trade with designs made for, and cards cut by, one of the above-named machines. Piano Card-cutting Machine The writer has not been able to trace the development of the piano card-cutting machine now in general use for cutting from the design. It has been employed upwards of forty years, and during that time it has undergone various modifications. It consists of an iron table, from the forward end of R 242 MECHANISM OF WE A VI NG PART which two uprights rise to support a pattern board 52'' long by 20'' deep ; the design is pinned upon it, and two straight-edges traverse it from side to side ; they serve as a guide for the eye in reading along the horizontal lines of the design. Both are moved up and down by a screw working in a nut on each end. The chief feature of a piano is its head-stock a (Figs. Fig. 135. 135 to 137); the lower part, consisting of two plates, is firmly bolted upon the table, with a space between them in the middle wide enough and deep enough to receive a card; the lower one contains a perforated cutting plate, and the upper one a guide plate for punches c. Two holes are bored through them near opposite ends for spindles h to work in; the latter support the movable head-piece a, in which a row of 12 vertical punches c, ^' from top to shoulder, and from shoulder to bottom, are fitted all VIII CARD-CUTTING 243 equal in diameter, and gauge to the holes in a Jacquard cylinder. Immediately in front, and exactly midway between them, the peg-hole punch d is placed ; it equals c in length, but is in diameter. The upper edge of guide plate e supports all punches c, by their shoulders; 13 keys, numbered from 1 to 13, have square shanks and large oblong heads ; all go through one of two slotted plates /, a spiral is threaded upon each, and a steel pin is pushed i2 iiXh%\ Fig. 136. through the shanks to hold the springs away from the outer edges of /. Every key head covers a punch when pressed in, 1, 2, and 13 are controlled by the right thumb, 3 by the little finger of the right hand, 4 to 10 by the remaining fingers of both hands, 1 1 and 1 2 by the left thumb. When pressure is removed from any key its spring pushes it out to clear the punch. The lower ends of spindles h are secured by lock nuts to cross head g (Fig. 135). A pin A goes through a forked pendant from cj and lever i, the latter is fulcrumed at y, and is fastened by pins Iz^ I to a three-armed lever m, that moves 244 MECHANISM OF WE A VING TART round centre n ; from the two remaining arms connecting rods 0 pass to treadles 'p which vibrate on pin ^. The above-named connections are made in such a manner that, as the card-cutter sits in front of the head- stock, with his feet upon the treadles, a downward move- A \ [D CD □ m m B J ji B C - - ■ w C - 11 ii 1; , ) \ 1 i Fig. 137. ment of his left foot will elevate head-stock a, and move, by means of plate all punches c out of the gap between the two fixed plates, then, if a card is pushed between them and some of the keys pressed in, a downward movement of the right foot will carry a down and force all punches that have their heads covered through the card ; those uncovered ^^^^ VIII CARD-CUTTING 245 will rest by their own weight upon it, but will not per- forate it. Down the middle of the table two smooth rails are placed for the wheels of carriage t (Figs. 135 and 138) to run upon, r is moved by a rack 5, a catch ^, and a weight u. Rack s is composed of stout pins driven into holes drilled in a metal plate that is screwed along one side of carriage r ; the gauge exactly equals the distance from hole to hole (n) n • nBBunyuuiiyiiyiyiiiiiiHBuniu nnmnnnnnnr t4 ff \\\\\ \ 3^ 140 Fig. 138. in a card. The shank of compound catch t passes through the table top and has a helical spring fixed on it by a pin to hold it down ; two catches move in the rack of pins, one is fixed to the catch box, but the lower one is free to slide ; it is kept in advance of the former by a spring, and the whole box is vertically moved when the left treadle is depressed by lever v upon which t rests, and a rod w attached to lever i. As t rises, the sliding catch comes into contact with the teeth of rack 5, weight u pulls back its spring, and carriage r recedes one tooth from the head-stock. 246 MECHANISM OF WE A VING PART In front of carriage r nipper jaws x are placed to receive a card which is pushed along a guide between the fixed plates of head-stock a, and below all punches c the nipper jaws are opened by depressing y with the left hand, the card is pushed close to a stop and level with the guide plates, the jaws of x close upon it and pull the card with the carriage as the latter slides tooth by tooth through catch i. One bar from the design is read, and one short row of holes is cut at each downward movement of the right treadle, also one pin in rack s is passed as the lifting treadle is depressed. An index cord is tied upon arm 14, led over guide pulleys along the pattern board, and has a small weight tied to the opposite end. One row of a fully perforated card is numbered above the holes pro- gressively and nailed upon the board ; when carriage r is close to head-stock (/, a knot is tied upon the cord, exactly opposite the first hole in the index card, and as the cutter is working, this knot should always cover a number corre- sponding with a bar number marked on the straight-edge. Cards are numbered at one end progressively before cutting to prevent mistakes ; the number on a card and that on the design for picks must always be the same. The numbered end is the 26 side, and is first pushed between the nipper jaws of carriage r ; peg and lace holes are cut, then bars from 1 to 26 may be used for design, a bar is left between 26 and 27 for middle lace- holes, and bars from 27 to 51 are for pattern, but with 51 the last peg-hole is cut, and beyond that again the end lace-holes. Card-lacing Before cards are ready for the loom they must be laced into a chain ; this is still largely a hand process. VIII CARD-CUTTING 247 although several machines have been invented for the purpose. For hand -lacing a frame is required with a series of wood or metal pegs fastened in it at both sides to face each other, and at such a distance apart as will suit the width of card to be laced. Cards are placed upon it side by side in proper rotation and held in position by the pegs which pass through both large holes ; the lacing is next threaded 0 0 — ! \ ^ 1 IJ p ^ 0/ p ^ 1 0 ' ~H y i ^ u -1 ul Fig. 139. amongst the lace - holes with a needle, so that it will be crossed over from right to left between every pair of holes in one card, and also between two adjacent cards, then back again in the same order, as clearly shown in Fig. 139. The defects of this plan are : inequality in the tension of the lacing, an excessive number of knots (for there are rarely more than from 40 to 50 cards between knot and knot in every line of lacing), and the slow speed at which the work is performed. 248 MECHANISM OF WE A VING PART Lacing is of various kinds and is in different conditions when used. Cotton and linen twines, twisted like a rope, are used singly and twofold ; they are also used after being plaited to form narrow braid. In some instances lacing is soaped ; in others it is steeped in boiled linseed oil ; in others, again, it is used from the ball, as delivered. It should always be in such a condition that it will not vary greatly in length as the atmosphere becomes dry or moist. Lacing-Machines Inventors have during the last thirty years endeavoured to construct a satisfactory automatic lacing-machine on the principle of a compound sewing-machine. The writer's earliest experience of such a machine was gained about twenty-five years ago, when an attempt was made to sew the cards together at two or three points simultaneously, by forcing stout sewing needles through the cards and locking their threads on the under side by shuttle threads. The number of stitches in each card, the inequality of the tension on the twine, and the lack of flexibility com- bined to render the inventors' efforts fruitless. Since that time, however. Count Sparre, Stahlknecht, the Singer Com- pany, Messrs. Reid, Fisher, and Parkinson, and others, have laboured upon the problem with more or less success. Count Sparre seems to have been one of the first to make use of the ordinary lace-holes of a card and to pass the needle threads only through those holes, after which the shuttle threads are linked into needle threads. By this means machine -lacing does not greatly difl'er from hand -lacing in appearance, except that in the former the threads are twisted in one direction continually. Eeid, Fisher, and Parkinson have greatly improved the VIII CARD-CUTTING 249 details of a lacing-machine, and have produced one that bids fair to entirely supersede the hand process. With it, 900 to 1000 cards can be laced per hour, and from 400 to 600 without a knot; the tension also appears to be well main- tained. When lacing 400^ or GOO'^ cards, three lock-stitch sewing-machine heads are employed, to pass ordinary lacing through the usual lace-holes only. The cards are fed upon endless chains with small projecting pegs to receive the large holes, and chains and cards are drawn forward at a speed commensurate Avith the width of card to be laced. The Singer Comipany employ two tapes, which they place above and below the card, and stitch all together by forcing needles and thread through them six or eight times in the width of a 600"^ card. This system of lacing is firm, but very troublesome where cards have to be fre- quently altered by adding to, or taking from, their number, as, for instance, in the manufacture of bed and table covers. It is also difficult to remove and replace broken cards. Card-wiring Cards are suspended from cradles above the loom by straight wires that project about 1^ beyond both ends of the cards. These wires are placed over the gap between two cards and tied with w^axed band to prevent slipping. They should be tied on the face side of the cards in order to keep them from touching the cylinder, or they will give an uneven bed for the cards to rest upon. Some prefer to push the wires between the two lines of lacing before tying ; but, on the whole, it is probably better to adopt the first-named plan. The distance from wire to wire is determined by the vertical and horizontal space 250 MECHANISM OF WE A VING PART available at the loom. Some wires are only 1 2 cards apart, but 16 to 24 are more general. Card Cradles Card cradles consist of two pieces of curved metal secured beneath the Jacquard cylinder in such a manner that cards, as they fall from the latter, will pass inside them. But as the pieces of metal of which it is composed are not more than or ^' farther apart than the length of a card, the wires will be caught, and the cards prevented from, falling to the floor. The bend of a cradle should be such as will prevent cards from piling up where they drop from the cylinder. And as fresh cards are taken up from the rear of a cradle, those remaining should automatically slide down to take their places, and thus leave a free fall to descending cards. The sectional shape of a cradle is unimportant ; provided sufficient firmness and holding power are given, nothing further is required. PART IX LAPPET SHEDDING is a peculiar system of shedding designed to move whip or warp threads out of their longitudinal positions by bending them until they assume a transverse direction, and after lifting each over a pick of weft, it is fastened at ])oth ends of every horizontal line, but floats loosely between those points. Elaborate figures are beyond the range of lappets, still there are many small eff'ects that can be economically woven by them, such as detached spots. IX LAPPET SHEDDING 251 and narrow continuous figures running more or less into stripes. The system imitates embroidery, and permits of A^arious colours and patterns being produced simidtaneously on any fabric, but plain or gauze grounds are oftenest employed, because they are least liable to have warp threads disturbed by a side pull. A lappet loom only differs from one of the ordinary type in having the following additions and alterations — namely, a groove provided in the slay bottom immediately behind the race board, but before the reed, and wide enough to receive needle frames and pin frame. Lappet wheel, whip rolls, and supports are also added. Large cranks of from ^" to 1" sweep, together with a slightly longer stretch, are desirable. A separate whip roll is required for each needle frame. They are situated above, below, or in both posi- tions with relation to the warp beam, as may be found most convenient, and separately weighted by cords and small weights. The reed is moved back from the race board to ^' to leave room for needle and pin frames. It is secured at the bottom by a piece of sheet iron bent to form a semi- circular groove for the reed to fit in, and is bolted at both ends to the slay back ; then the slay cap front is bolted behind the swords. Figs. 140 and 141 illustrate an arrangement of parts designed to weave continuous lappet figures, or spots, in which no portion of a fabric is entirely free from figure throughout its length. Fig. 140 is a front elevation : a are the slay swords, h a bracket bolted on the front of a. It supports an adjust- able stud c, on which ratchet and lappet wheels d turn freely. Both are either cast in one piece, or if the lappet wheel is made of wood, are bolted together. The latter 252 MECHANISM OF WE A VING PART consists of a cylindrical drum having a series of varying indentations cut in its front edge in such a manner that, as the ratchet is rotated tooth by tooth, feeler e will be moved from a ridge to a hollow, or vice versa ; e is centred on a pin in midway between the points of contact with rf, /. It is flattened to a knife edge, where it touches the 2 1 2 \. Fig. 140. lappet wheel ; but its upper arm is rounded, and has a cord tied upon it that is carried to, and made fast upon, a short pendent bracket screwed on needle frame g. A spiral spring A, or a cylindrical piece of elastic, is hooked into pin frame shifter i at one end, and made fast to needle frame g at the other. The effect of A is to hold feeler e constantly against the irregular face of lappet wheel d; hence, as d turns round, e vibrates in unison with the IX LAPPET SHEDDING 253 ridges and hollows of and transfers the movement laterally to needle frame and through it to whip threads that enter the needle eyes. All the above-named parts are carried backward and forward by the swing of the slay, but the rotation of d comes from an adjusting catch, fixed in the end framing, and supported by a bracket. As swords a move forward with c/, the teeth of the latter are successively brought into contact with pawl 'p^ and rotation to the extent of one tooth ensues. Lappet wheel 6? is a hollow, built-up cylinder of close- grained wood, Y thick, which is filled in at one end to supply a means of fixing it upon the side of a ratchet wheel. Its periphery and front edge are turned true and 254 MECHANISM OF WE A VING PART made smooth, then a metal comb, with pointed steel teeth, equal in pitch to the reed to be used, is pressed against the cylinder as it turns in a lathe, and thus a series of parallel lines are scratched along its surface. After taking the cylinder out of the lathe its periphery must be divided into a number of equal spaces to suit the picks in one Fig. 142. repeat of the pattern to be woven, and lines drawn through each point, parallel with the cjdinder's axis. For example, let a (Fig. 142) be the design, h (Fig. 143) the periphery of the lappet cylinder when opened out, then thin vertical lines c are those scratched by the comb teeth. They are indefinite in number, but in pitch each equals the space occupied by two warp threads in the fabric. Thin horizontal lines d are parallel with the cylinder axis. Thirty -two horizontal spaces on design LAPPET SHEDDING 255 paper a give one repeat of the pattern, but as each line of whip must be fas- tened at both ends of every float, 64 picks will equal one revolution of the lappet wheel ; hence there are 64 spaces d. Thick lines e show the form of teeth required to reproduce pattern a on cloth. In the first horizontal space of a, and on the first line of 2 dents, or spaces, are passed in each case to find the starting-point of the figure, which at 1, a floats over 4 dents. It will be seen that 2, 6? is 4 dents lower than 1, d. Therefore feeler e (Fig. 140) will move the needle frame g through that space. On 2, a (Fig. 142), 4 dents are passed, and 4 taken for pattern. On 3, d, 4 ^dents are also passed over ; then on 4, d, 4 more dents are taken, and so in like manner up to 16, a, and 32, d, where the first spot ends. At 17, a, 18 dents are passed and 4 taken. At 33, 34, d, the same order is maintained. In many detached spots, such as a, it is necessary to add binding to each figure by moving whip threads across a single dent immediately before a spot begins and ends, for this has the same effect of preventing whip threads from being pulled out that fastening-ofF has In all such cases extra 501 401 20 B m 256 MECHANISM OF WE A VI NG PART teeth must be provided in the ratchet, and extra horizontal spaces in lappet wheels, for both equal one pick of weft in the piece. Great care and accuracy are essential to the proper cutting of a lappet wheel, else the pattern will be defective, but beyond these little more is needed. A needle frame g (Fig. 144) is made of wood l|''to If' deep by thick. Each frame has two horizontal grooves 4" to 5'' long, cut equidistant from the ends, and generally lined with brass, to fit freely upon flat pins 0, secured to pin frame j/*, in such a manner that a vertical lift can be imparted without any side movement. A number of steel o GO Fig. 144. or brass needles are driven into the wood and protrude 2^'' to 3|-''. They are all flattened at the top, pointed, punched, and made perfectly smooth, to form eyes for the whip threads to pass through. The exact number and positions of the needles are determined by the size of the pattern and the number of times it is repeated on the width of the fabric. Thus pattern a (Fig. 142) has 32 dents to a repeat. If the reed has 20 dents per inch, and 37'' of it are filled, 37 X 20 = 740 dents occupied by warp. 740 -f- 32 dents per pattern = 23 patterns. 37'' of a frame is divided into twenty-three equal parts, and a needle is driven in at each mark. Every frame has two movements — namely, a vertical one to lift whip threads above a moving shuttle. IX LAPPET SHEDDING 257 and a lateral one which bends whip threads into a transverse position, and thus forms a pattern. A pin frame 7 (Fig. 140) is merely a false reed that serves as a background for the shuttle to run against. In form it is not unlike a needle frame, but the pins are long by ^' in. diameter. They are pointed at the top, and flattened at the front from tip to base. They are Y apart throughout, and driven into the frame so that, when lifted, all will touch the race board and form a true guide for the shuttle. Both ends of this frame are tipped with brass, and slide in grooved brackets ^, which are screwed to the slay bottom. Each rises above, and sinks below, the point of attachment to the slay. A horizontal shaft I vibrates in bearings bolted on brackets h. It carries two pulleys and one pulley n. The former have straps 8'' long by Y' broad, fastened on their surfaces by set screws, and the otherwise free ends are secured to vertical arms % pendent from pin frame /. Pulley n has a similar strap %)" long fastened upon it, but wound in an opposite direction to those on m, whence it is led over a guide pulley, and finally bolted to the front rail of the loom. The effect of this arrangement is, that as shaft I moves back with swords the strap on n will unwind, and in doing so will wind the straps of m upon the surfaces of their respective rollers, and push up pendants i, pin frame j, and needle frames g, A spiral spring, acting through cord connections upon pendants % assists in drawing down all the frames as swords a move forward. A frame must invariably be lifted high enough to allow the shuttle to pass under whip threads when the loom cranks are on the bottom centres. With this arrangement, however, it will continue to rise until the cranks are at their back centres. Such additional rise s 258 MECHANISM OF WE A VING TART is superfluous, and puts undesirable strain upon the threads. The following plan for moving needle and pin frames is preferable. Bolt the strap from to a tappet treadle that rests by its own weight upon a cam, and let the full part of the cam unwind strap from n and push up pendants i. By this contrivance frames are only lifted to the height required, and a minimum of strain is put upon whip threads. Two or more patterns can be woven simultaneously, but a separate lappet wheel is required for each. For two patterns, another wheel larger or smaller than d is situ- ated in a corresponding space on the opposite side of the loom, and connected in the manner already described. Where both patterns repeat on the same number of picks, and where one is a multiple of the other, two cylin- ders can be secured to a single ratchet wheel, one inside the other; then by employing two horizontal instead of vertical feelers 6, each set to be acted upon by different cylinders and connected with separate needle frames, mechanism is not only economised, but a perfect move- ment is given to the needle frames. A horizontal feeler is invariably superior to a vertical one, because the latter must move in the arc of a circle, and, as a consequence, will carry difl"erent parts of its sur- face into contact with the lappet wheel, and to a small extent an inaccurate traverse will be imparted to the needle frame. It is also usual to employ an elbow lever and a cam to withdraw all sliding feelers from the sur- faces of lappet wheels before rotation commences, and thus much friction and wear are saved. Instead of using a cylindrical drum, such as d (Fig. 1 40), a disc, drilled at regular distances with holes arranged to IX LAPPET SHEDDING 259 form one or more circles, may be substituted, and one or more patterns will be formed by securing pins of various lengths in the holes. Such a wheel has the great advan- tage of permitting a large number of different designs to be woven by simply readjusting the pins. Also, that one disc can be driven at different speeds to accommodate patterns of different lengths, provided the disc holes are a multiple of the design. One of the most important points to be attended to in lappet-weaving is the regulation of tension upon whip threads, and very delicately adjusted wires are provided for this purpose. Fig. 145. The threads from whip rolls are led between spring cords and healds under the reed and through the needle eyes. Spring cords a (Fig. 145) consist of two wooden end- pieces y^' thick by ^' long and \" wide, into which two wires that are long enough to reach across the warp, are driven, and secured apart and parallel. Two holes are drilled in both end-pieces f on each side of their centres, and cords c, d are threaded through them, twisted, and tied upon one of two flat springs that protrude from the loom framing. Then, as the whip threads pass in front of the upper and behind the lower wires ^, it follows that the degree to which cords c, d are twisted determines their power to readily give off or take up whip in unison with 26o MECHANISM OF WE A VING PART the rise and fall of the needles. It is a nice point to adjust tension so as to prevent strain and still hold the threads tight. The foregoing lappet mechanism is not in such exten- sive use as the Scotch type ; nevertheless, it is well adapted to the manufacture of continuous patterns, and has one very important advantage to recommend it — namely, that every added part is inside an ordinary loom framing, and consequently no additional floor-space is needed. The essential features of a Scotch lappet loom are : That needle frames are moved vertically in, and horizontally by, shifter frames ; that a large wooden wheel, with a groove cut in one side to receive a feeler or peck, is rotated through a space of one tooth on alternate picks ; that shifters are moved ^'^-^^^^^'^ by the dead weight of a lever hanging first on one, then on its other end, and through a space equal to the width of slot in the lappet wheel ; also, that all upward movement comes from cams. Such a lappet wheel is situated on one side of a loom beyond the fabric, and swings with the slay. It has an irregular groove cut in one of its sides to limit the lateral movement of a needle frame. This wheel is of wood ; sycamore, or some other close-grained hard timber is preferable. Its diameter is not of great importance, pro- vided it does not revolve so rapidly that the groove becomes worn by frequency of contact with the peck at any one place. IX LAPPET SHEDDING After the wood has been turned to the size and thick- ness required, a comb with a pitch, generally equal to half that of the reed, is pressed against the revolving disc to describe a series of concentric circles on one of its sides. The circumference is next divided by radial lines into as many equal parts as there are horizontal spaces on the Fig. 146 (6). design (see Figs. 146, in which a is the design, h the lappet disc, c concentric circles, 16 radial lines corre- sponding with the 1 6 horizontal spaces of design a). Each space betw^een 2 radial lines equals one tooth in h and two picks of weft in the fabric. A vertical space in a represents two, and a circular space c four warp threads. By comparing c, it will be found that number 1 and following numbers agree in space between extreme 262 MECHANISM OF WEAVING PART points, after a fixed allowance has been made at c of four threads for the thickness of the peck. A groove must be cut in h to the exact dimensions of the thick marks. Two patterns can be woven from one wheel by cut- ting a second groove inside or outside the first one as shown, but beyond this a new wheel is needed for a new pattern. A wheel h has a tooth cut on each space between 2 radial lines c?, and it is negatively turned by a flat iron catch that is grooved to permit the inside edge of h to pass through, then as the solid upper part of this catch comes into contact with the wheel on alternate picks, a forward movement of one tooth takes place immediatel}^ the cams employed to slide the needle frames are midway in their lift, because at that point all strain is taken from the peck. Movement is given to the catch by a bowl which works on one side of a needle frame cam and depresses a lever ; the latter acts through a spiral spring upon the catch ; the spring is employed as it expands in case of obstruction, and prevents a smash. When weaving intermittent lappet figures, a sliding connection is made on this lever that enables it to be stopped and started at pleasure. A horizontal traverse is given to a needle frame by means of two cams fitted on the extreme ends of the tappet shaft in such a manner that when the full side of one is at the top the full side of the other is at the bottom. Two small levers rest by their own weight on the cams, and from their forward ends straps are taken over guide-pulleys and made fast to opposite ends of a shifter. The latter is a frame, with two upright arms grooved on the inside for the ends of a brass-tipped needle frame to fit in and slide IX LAPPET SHEDDING 263 up and down without any end movement. A shifter is connected by an adjusting piece to the peck that works in the groove of a lappet wheel. A shifter lever must be sufficiently heavy to move a frame horizontally when the thin part of a cam is uppermost, but the width of groove in the lappet wheel determines how far move- ment shall be carried, for a lever can only cause a shifter to slide so long as the peck has space to move in ; immediately the latter touches one side of the groove, the weight of a shifter lever is sustained by the shifter, until at half the lift of a cam it is removed. Lateral movement begins when all needle points are below the bottom line of warp, and the reed in moving forward is approximately J'' from the fell of cloth ; it must end before the vertical movement begins — namely, before the reed has traversed in its backward direction from the beating-up point, and the needles should be fully lifted Avhen the pick is delivered. The consumption of whip always largely exceeds the length of cloth woven, but the exact amount depends entirely upon the pattern. To estimate the length of whip required for a given length of fabric, the design must be counted to find how many dents the figure covers at each traverse of the needles ; and when the sum of all movement is obtained — The number of dents per pattern x the needles in a frame x the repeats of pattern in one yards of whip yard of fabric _ for one yard dents per inch in reed x inches of cloth, per yard Take as an example, a (Fig. 142), in which there are 32 horizontal lines of whip, 23 needles in a frame, 20 dents 264 MECHANISM OF WE A VING PART per inch in the reed, and 40 picks of weft per inch in the fabric. The lines of whip vary in length as follows :— Number 1 = 4 dents. 2 = 4 3 = 4 4 = 4 5 = 6 6 = 7 7 = 8 8 = 9 9 = 9 10 = 8 11 = 7 12 = 6 13 = 4 14 = 4 15 = 4 16 = 4 Total 92 dents in the first half of the pattern, and an equal number in the second half. To these must be added 10 dents passed over in moving from the first to the second spot, and 22 dents in moving from the end of number 2 spot to the beginning of the first spot in the next repeat. Altogether, 92 + 92 + 10 + 22 = 216 dents. To find the number of repeats of pattern per yard of cloth — 40 picks per inch x 36" per yard ^ ^ — ^ ^- = 221 patterns. 64 picks per pattern ^ 216 dents x 23 needles x 22^ repeats per yard 155^ yards of whip for 20 dents per inch x 36" ^ oi^e yai'd of cloth, per yard X PICKING 265 PART X PICKING Picking mechanism is constructed and timed to follow shedding, and the operation consists in passing a shuttle containing weft between the upper and lower lines of warp. With the exception of a few experimental looms, the method adopted during many centuries by weavers of all countries was to take the shuttle in one hand, throw it through the shed, and catch it with the other hand as it emerged from the opposite side ; but in 1738 Kay, of Bury, introduced what was known as the " fly shuttle " ; it was simply an addition to an ordinary slay of boxes placed at opposite ends for the reception of a shuttle. A box consisted of a bottom, two sides, and one end. Over each box, and extending its entire length, a metal spindle was fixed, having a diameter of about f", for the purpose of guiding a picker or driver employed to propel the shuttle. Both drivers were connected by cords to a wooden handle known as the picking stick, which the weaver grasped with his right hand, and by making a rapid lateral movement of his arm, the shuttle was jerked with sufficient force to ensure it entering the opposite box. Kay's invention was slow to find favour with weavers of his own time, but when once adopted, it soon displaced the older method, and became almost universal. Constructors of modern looms have principally confined their attention to improving and altering Kay's invention to suit the new conditions which were introduced with steam as motive power. It is indeed difficult to imagine a 266 MECHANISM OF WE A VING PART better or simpler all-round motion ; still, with these advan- tages, picking remains the weak point in a loom. Experi- ence has proved it to be uncertain in action, costly to keep in order, and by far the most dangerous part of the machine. More serious accidents result from defective picking than from all the remaining parts of a loom put together; notwithstanding the fact that throughout an entire century inventors have laboured to remove the most objectionable features ; their efforts have entailed the spend- ing of large sums of money in developing and patenting so-called improvements and safeguards to prevent accidents to work-people, breakages of machinery and warp. These inventions include a host of picking motions, shuttle guards, swells, check straps, fast and loose reed appliances, pickers, and an endless variety of details con- nected w^ith picking, many of which are evidently the result of a misconception of the problem, for they attempt to deal with the effect instead of the cause. To fully appreciate the defects that are liable to be de- veloped by a negative propelling motion, attention must first be directed to the shuttle itself, which is made from hard, smooth wood, pointed at both ends, and tipped with steel ; it is hollowed out in the centre for the reception of weft, the latter being used in the form of a cop, wound upon a wooden pirn or a paper tube, and pressed upon a metal tongue in the shuttle, that is hinged at one end and extends almost the entire length of the hollow part. This tongue retains the weft in one position during the operation of weaving, and allows it to be drawn away as a continuous thread through an eye fixed in the front of the shuttle. The rapidity of motion in a shuttle, together with an imperfect controlling force, is the chief cause of serious defects in the working of ordinary picking motions ; but X PICKING 267 it remains for us to analyse the various forces which tend to divert a shuttle from its true course. Weight affects its movement, for a study of mechanics teaches us that the energy possessed by a moving body varies in proportion to its weight and the square of its velocity ; hence, as a shuttle loses weight each time it moves across a loom equal to the weight of weft drawn away, a proportionate diminution of force takes place, and any obstruction is more liable to divert a light than a heavy shuttle. Therefore, unless an almost empty shuttle is heavy enough to overcome every resistance, and pull its weft close to the selvage of a fabric, there will be a constant risk of flying out, and this is the evil most feared by weavers. The drag of a coarse weft as it passes off at the shuttle eye being greater than that of a fine thread, necessitates a change from a light to a heavy shuttle when an alteration is made from a thin to a thick weft. The position of the centre of gravity with relation to the direction of force employed to propel a shuttle has an effect upon its motion. The centre of gravity has been defined as a point which does not change its position in a body when that body is turned in any direction; and further, if a body free to move in any way is subjected to a blow, motion will be in the direction of the straight line in which the blow is delivered, provided the line passes through the centre of gravity ; but unless the force acts through the centre of gravity, the body will not merely move as a whole, but it will revolve. A shuttle is not free to move in any direction, still it moves at a high velocity without being under positive control, and it is therefore imperative that the force em- ployed to move it, and the parts used as guides, shall 268 MECHANISM OF WE A VING PART act to minimise any tendency to rotate, or, in other words, to fly out. For the moment let it be assumed that a shuttle is a parallelogram, having its centre of gravity at the centre of the body (see Fig. 147). If force d passes through the centre of gravity c, and parallel to all sides, two of which A -D B Fig. 147. are shown at a, movement will be straight without the aid of guides ; but the slightest obstruction, such as rough, knotty, or entangled warp, will be sufficient to set up a rotary motion. If the line of force does not pass through the centre of gravity, then the distance between centre of gravity and line of force represents the leverage which will be exerted to produce rotation (see Fig. 148), where a, h represent two D Fig. 148. sides of a rectangle, c its centre of gravity, d the line of force, and c, e leverage. Eotation can be checked by placing guides, say reed and race board, against those sides of the rectangle farthest from the line of force ; for this reason some textile experts advocate placing the shuttle tips a little nearer the top and front than the back and bottom ; or so con- structing a shuttle that the line of force shall be above and before the centre of gravity. By so doing it is argued that PICKING 269 pressure will be exerted against reed and race board, and the tendency to fly up or forward will be checked at a slightly increased cost of friction and power ; the latter are small matters comxpared with the risks attending a shuttle flying out. Other experts advocate placing a shuttle tip as near the race board as possible, in order that it may pass beneath entangled warp, and thus prevent an upward movement in the shuttle. Both points are worthy of consideration. It has been assumed that the centre of gravity in an ordinary shuttle occupies a fixed position, but careful con- FiG. 149. sideration of the problem will lead us to the conclusion that it is a constantly changing point, for if when a shuttle is loaded with weft its centre of gravity is at the centre of its mass, every time it is driven across the loom that point is moved slightly back in proportion to the weight of weft drawn away, because weft is invariably pulled from the forward end of a cop ; hence, as weaving proceeds, the front of a shuttle becomes lighter, but the back is unaltered until the cop is about half drawn off. The alteration may not be great, still it is one of the many small things that tend to make the whole system of picking uncertain. When weft is passing through the eye, a side pull of varying intensity is exerted upon the shuttle, this also 270 MECHANISM OF WEAVING PART has a tendency to pull it out of a straight line. By a reference to Figs. 149 and 150, in which the arrows indi- cate the direction of motion, a the fell of cloth, h the reed against which shuttle c moves, d the weft rimning diagon- ally between c and a, it will be obvious that any obstruc- tion to the free passage of weft through the shuttle eye will, in Fig. 149, tend to draw the front of c from and cause it to fly out; whilst in Fig. 150 the forward end of c will be pressed against and slightly reduce the above-named tendency. This is due to placing the eye of B A Fig. 150. c out of the centre, but the efl"ect in either case will be in direct proportion to the intensity of such pull. It is well known that weft is less free to pass from a shuttle eye when a cop or pirn has nearly given out than when full, which is accounted for by the presence of a coarsely-pitched coil of weft found between the unused cop and the tip of the shuttle tongue in the former, and the absence of coils in the latter. Every coil puts addi- tional tension upon the weft, and as a result we find the straight motion of a shuttle modified in a twofold degree : first, by a reduction in weight, which results in a reduc- tion of force ; and secondly, by the increased power of the diagonal pull being exerted when the shuttle is least cap- able of resisting it. PICKING 271 In the next place, the actual movement of translation must be considered. It has been previously mentioned that the two guides of a shuttle are the race board, upon which it rests, and the reed, against which it presses ; both forming parts of the slay, and swinging on a centre by the action of cranks and connecting arms. It follows, therefore, that a shuttle partakes of this swinging motion in addition to its own of translation. Fig. 151 is a diagrammatic representation of the above- named parts : a is the shuttle, h the slay bottom, c the reed, d a sword which moves partly round the centre of rocking-shaft ^, / a connecting arm, g a crank, li a dotted line showing the circular path of g. When a shuttle begins to move, the various parts occupy positions corresponding with the solid lines ; the slay is in motion, and continues to be pulled back until crank g reaches point 2 in the periphery of circle li^ where a slight pause takes place for reasons that are fully dealt with in Part XV. The position of each part at this time is indicated by the same letter, with the addition of a dash against dotted lines. It will also be noticed that when the sword is at d! ^ all the parts are in a lower plane than at d ; the fall of the slay bottom being indicated by the space between lines 4 and 5. In consequence of which, during the movement of slay from d to d\ the shuttle has been travelling across the loom, moving back, and falling with the slay; but when crank g moves round point 2, the shuttle simply continues to pass across the loom. After leaving that point it moves up and forward with the slay until it finally reaches the opposite shuttle box. Fig. 152 will further elucidate the aggregate motion of a shuttle. Let line a represent the width of loom, 272 MECHANISM OF WEAVING PART d the backward movement of slay ; it will then be found that a shuttle, in passing from one side of a loom to the other, moves diagonally from 1 to 2, straight from 2 to 3, and in the opposite diagonal from 3 to 4. But this does not show all its motions, for the fall and rise are not taken into account. In Fig. 153, a equals X PICKING width of loom, 4 and 5 the fall of slay. The shuttle falls from 1 to 2, moves straight from 2 to 3, and rises from 3 to 4. Add to the above-named modifying influences that a shuttle begins its journey when a slay is moving back at nearly the same speed as a point in the periphery of circle S 3 Fig. 152. /i, (Fig. 151); but as the crank approaches point 2, the velocity of slay is rapidly reduced, until at 2 a pause is reached, then from 2 to 3 an increase in velocity propor- tional to the decrease from 1 to 2 is made. A shuttle, therefore, partakes of many variations in velocity and direction, but they do not all necessarily tend to throw out the shuttle. On the contrary, some of them may assist in reducing such a tendency. For example, when a shuttle begins to move across, it is also travelling back with the slay at its greatest velo- ! — ^ —4 5 2 A 3 5 Fig. 153. city, and if at such times reed and box back were removed, the shuttle would continue to travel in the same direction ; but the rapid checking of the slay's speed causes the shuttle to press against the reed with a force proportionate to its weight and acquired momentum, and this to some extent minimises its liability to fly out. If due consideration is given to these points, it will 274 MECHANISM OF WEAVING PART reveal the existence of many things that tend to render motions for throwing a shuttle less efficient than is gener- ally supposed, and it will become obvious that moving such an apparently simple article as a shuttle presents a complex problem for solution. Before attempting to describe the varied mechanism used to propel a shuttle, it will be best to first determine what features may be considered as essential to a good motion, for then the relative values of such parts can be more readily estimated. In the absence of a better definition, the following points are suggested as essential to a good pick : — (1) Power consumed. When a shuttle is negatively driven, an enormous waste of power results, partly on account of the impossibility of accurately gauging the force required, and partly because the best of motors is liable to variations in speed. A shuttle must never be permitted to rebound after reaching a shuttle box, for which reason swells are em- ployed in such a manner that it has been affirmed that the force required to drive a shuttle into or out of a shuttle box is equal to that required for driving it through a shed. On this assumption three times the actual power required for useful work performed is taken from the engine, and twice the necessary power must be created by springs or other appliances. For example, let the force required to pass a shuttle across a loom be represented by one unit of work ; then to drive it out of number 1 box a second unit is employed, and to drive it into number 2 box a third unit must be provided, hence 2 units are wasted to 1 profitably employed. These 2 units are spent in annihilating 2 other units stored in brakes or springs, consequently 5 units of X PICKING 275 work are consumed, and 4 of them are wasted each time a pick is delivered. Even assuming that more force is required to drive a shuttle through the warp than to drive it into or out of a box, the force wasted still largely exceeds that usefully employed; and in a weaving shed containing hundreds of looms the waste is simply enor- mous. Therefore the first proposition is : that as weight of shuttle and time taken to move it are the main factors, force in excess of that required to pass a shuttle across in the time allotted should not be employed. Fig. 154 illustrates one of the many attempts that have been made to reduce this waste of power, and also shows in what direction inventors have sought for improvement ; it is a contrivance for taking pressure from the shuttle on leaving a box : a is the swell lever, to one end of which 276 MECHANISM OF WE A VING PART strap h is attached, and its opposite end is bolted upon connecting arm c / is a crank and e a slay sword. As arm c is depressed by the rotation of crank rf, strap h tightens and pulls back swell a, thereby relieving the pressure upon the shuttle. If such an appliance can be considered as satisfactory, the ultimate saving of power cannot be great ; it is repre- sented by the difference in leverage between the points of contact with the swell lever a ; the shuttle acts at the lower, and the strap at the higher point ; beyond this it merely results in diminishing the power consumed by the pick and increasing that of the slay. (2) A desirable motion. The movement of a negatively-driven shuttle is essen- tially a jerky one, and frequently produces the most dis- astrous results ; for its velocity is developed suddenly, and is greatest where most defects are likely to be found ; that is, as it enters the shed, which at this time, if fully open, will, as the reed moves back, give considerably more room for the shuttle to pass, and granting that obstructions are less likely to throw a shuttle out at this place than at others, there is still a great tendency to break the warp. Again, the power required to develop a high speed sud- denl}^ is greatly in excess of that required to develop it gradually. The second proposition is, therefore, that a shuttle should begin to move slowly, and develop speed up to the slay centre ; then, from that point, a corresponding decrease should take place, until a final pause is reached at the opposite side. (3) A positive motion. Nearly all the serious accidents that occur in a weav- ing shed result from shuttles flying out ; this is due PICKING 277 entirely to negative picking; and as the preceding pro- position necessitates the application of a positive motion, the results of which would be the absolute safety of work- people, a considerable saving in power, a great reduction in breakages, and a consequent cheapening of production. The third proposition is that a shuttle should be under such complete control throughout its entire movement that loom and shuttle would start together, irrespective of the position of the latter when the former was brought to a stand. (4) Altering the speed of a loom. All weavers know that if the speed of a loom is variable, the power of the pick is altered to such an extent as to cause the loom to knock off ; if its speed is accelerated the shuttle rebounds, through excessive force ; and if retarded, it does not reach the opposite box, owing to insufficient force. A similar effect is noticed if two shuttles of unequal weight are used in a loom, when, if the pick is correctly set for the light shuttle, the heavy one will be driven with violence against the opposite box end ; but if the pick is set for the heavy shuttle, the light one will barely reach its destination. As this is a serious and one of the most obvious defects of negative driving, the fourth proposition is that picking mechanism should be so constructed that force remains constant, no matter how the speed of loom is altered. (5) Principal motions connected. Smashes are of frequent occurrence through shedding, picking, or box motions working independently of each other ; to obviate which is the object of the fifth proposi- tion — namely, that positive connections should be made to prevent one from getting out of time or rotation with the other. 278 MECHANISM OF WE A VING PART The writer is fully conscious that few picking motions now in use are capable of accomplishing the task demanded on the preceding pages, but the student must seriously consider whether a piece of mechanism which gives the shuttle a blow and then leaves it without further control is true in principle or economical to work. Also to what extent the application of a guard to prevent such a shuttle from flying out is an attempt to deal with the effect instead of the cause. The mere enumeration of the requirements of a picking motion is sufficient to show that a formidable problem must be solved before a truly satisfactory pick is obtained ; and an analysis of the defects arising from inattention to the points already mentioned will reveal the pressing need for improvement. Attention has been almost constantly directed to this subject, and numerous motions have been put upon the market from time to time, some of which are negative, others are positive in action. The former are of three kinds — namely, under, over, and pick and pick motions ; the difference between an over and an under pick mainly relates to the position occupied by the picking arm fulcrum which, if entirely below the shuttle boxes, is known as an under- pick, but when some portion of it projects above the boxes the motion becomes an over-pick. Of under-picks there are, a, those that deliver the blow from the bottom shaft, and J, those that deliver it from the crank shaft. In most of the motions included in division a it is possible to shape the piece struck to impart more of a push than a blow to the shuttle, and when this can be satisfactorily accom- plished the best results are obtained ; it must, however, be admitted that less attention has been given to this matter than it deserves, and as a consequence, it may be generally PICKING 279 X affirmed that such under-picks when compared with over- picks consume more power, work less smoothly, and the risks of shuttles flying out are greater. Jn_C o o yio/'^~^ M Fig. 155. Pick and pick motions are applied to looms that have a series of boxes at each end, and where tAvo or more shuttles require to be driven from one side of a loom before any are returned from the opposite side. They allow a single 28o MECHANISM OF WEAVING PART pick of any colour of weft to be passed through a shed which is rarely attempted with alternate picking. Most of them are modifications of alternate picks, and can be con- verted without making important changes at small cost, but others require the addition of numerous parts. The lever pick is one of the best known under motions; it consists of an arm a (Figs. 155 and 156) fastened upon the bottom shaft, to carry an adjustable stud and bowl J. At one part in the rotation of a, bowl h comes into contact with a curved metal plate c, bolted to a wooden lever that is fulcrumed outside, and near the back of the end framing ; at its forward end it rests upon an arm 6, pro- jecting from shoe / of the picking arm g ; this arm has its fulcrum pin fixed in a bracket on the rocking- shaft and swings with the slay, g passes through a slot in the shuttle box bottom and has a leather picker dropped over its upper end ; a wooden or iron rib on the top of a shuttle box prevents the picker from flying off. When bowl h strikes the picking plate c lever d is depressed ; the picking arm and picker are tilted over, and H Fig. 156. PICKING 281 the shuttle is driven across ; spring and strap li pull l)ack arm g to its normal position. It will be understood that duplicate parts are placed at the opposite end of the loom to drive the shuttle back again, also that when the pick is acting at one side it is inoperative at the other. If a loom has two or more shuttles, picking arm g passes through a slot in front of the boxes, a buffalo- hide picker is dropped over it, and a guide spindle is pushed through the picker. A further alteration is made in picking plate c, with a view of preventing shuttles from being thrown out of the boxes when the cranks are turned backward. It acts as follows : — d is the lever, d a plate bolted upon it, c a sliding picking plate through which bolt ] passes, and slot h serves as a guide for c. When the loom is running in its ordinary direction, plate c is pulled by helical spring I against a fixed shoulder m, and the pick is delivered in the usual manner ; but when the loom is reversed, the bowl strikes the full side of c, distends spring and plate c slides along the surface of d without depressing lever and consequently without moving either picking arm or shuttles. It is of the utmost importance that picking plate c shall be correctly curved, but before this can be done it is necessary to fix upon the length of lever d^ the relative positions of picking shaft and lever centres, the time allowed for picking, the depression of lever d at its point of contact with bowl and the ratio of depression. If the following dimensions are assumed : — namely, centre of picking shaft to bowl centre 4'', bowl radius, centre of shaft to centre of lever pin 15'', — the time for delivering the blow usually varies from \ to of a pick ; -^^ of a MECHANISM OF WE A VING PART i i I revolution of picking shaft will be taken here ; thick- ness and length of picking plate \" by ^" ; length and depth of lever d 36" by 2|"; depression of lever at point of contact depends partly upon length of shuttle box, partly upon position and form of connection with shoe of picking arm, say X'\ ratio of depression 1, 2, 3, 4, in equal times because lever d must begin to move slowly, and rapidly in- crease in velocity until the end of a pick is reached. We are now in a posi- tion to proceed with the construction of a picking plate. Fix the position of shaft centre a (Fig. 157), describe a circle with a radius of ^' to trace the path of bowl centre. Drop a vertical diameter line 2, 3, and from 3 divide circle 1 into sixteen ecpal parts ; also divide space 3, 4 into four equal parts. Eound point 4 describe bowl circle 5, 1|" radius ; draw the upper line 5 of picking PICKING 283 plate c at right angles to line 2, 3, and touching the periphery of bowl circle. Show the thickness of plate by making a second line \" below and parallel with 5. Next find the centre line of lever d and fix pin centre X by producing a line at right angles to 2, 3, and half the depth of d below line 6, =2|''-f2 = lf^ To find centre X, continue line 2, 3 to the centre line and mark off 15'' to the right, and 21" to the left will give d! the extremity of lever ; complete it. With X 3 as radius describe arc 8, and from the same centre, through each point formed by a radial line in division 3, 4 intersecting circle 1, describe arcs 9, 10, 11. From point 3 on arc 8 cut X' to equal the fall of lever d and divide it into ten equal parts ; then, with a as centre, describe concentric arcs cutting points 1, 3, 6, 10, to give the ratio of depression. From 4 take as bowl centres each point in succession, where the arcs, concentric with a, cut arcs 11, 10, 9, 8, and describe corresponding circles. Trace a line touching the periphery of each, and it gives the correct curve, cut off the curve at a point 2" above lever d^ drop perpendicular 12, draw any curve on the opposite side, then make plate c ^" long by taking 3'' on each side of 12, and the picking plate is complete. The lesser pick is much used as a pick and pick motion for silk weaving, because it is clean, simple, easily altered, and capable of throwing shuttles in a very irregular order. Figs. 158, 159, and 160 give respectively plan, front and back elevations of this motion. Two arms, both furnished with two bowls, are keyed upon opposite ends of the bottom shaft, and either can be made to strike or miss one of the picking plates c any number of times in succession. The latter are pivoted at and the rear end of each is connected by bar 0 in such a manner that one is resting on and parallel with lever a, whilst the other is moved from 284 MECHANISM OF WEA VING PART that part of its lever where a bowl would otherwise strike it. The mechanism for governing picking plates c consists of a slide wheel ^, which is set-screwed on the bottom shaft, and has on one of its sides a bead with two breaks exactly opposite each other ; in the centre of each break a stud is fixed to come into contact with a notch of an eight-sided star wheel ^, and cause q and a chain barrel, of which it -15] ^ Fig. 158. forms a part, to partially rotate on a stud inside the framing. Chain r is built up of thick and thin links to suit the required order of picking; it is passed round and turns with the barrel one link every pick. Lever 6-, also centred inside the framing, carries bowl i ; the latter is pulled upon chain r by a spiral spring and rises or falls as the link beneath it is thick or thin. Bell crank lever v moves on a stud in the back rail, and is connected by a link to lever s. A stud connects v with X PICKING 285 bar 0, therefore, as lever s moves up and down, one picking plate c is pushed out of striking position, and the other Fig. 159. plate pulled under a bowl. Any desired order of throwing the shuttles can be attained by laying the chain to suit it. Figs. 161 and 162 show respectively plan and elevation of Jackson's pick, which also delivers its blow from the bottom shaft a by an arm having at its outer end an antifriction roller c, that is carried by the rotation of shaft 286 MECHANISM OF WE A VING PART a against the under side of picking finger d ; the latter is keyed upon a short shaft e placed at right angles to and with its axis in the same horizontal plane as shaft a ; e is Fig. IGl. Fig. 162. supported in brackets /, which are bolted to the end framing ; it also carries a long curved arm having a connection made with picking arm A, by bending a strap i round the shoe of A, and bolting a second strap j to both i and g. PICKING 287 When bowl c pushes up the picking finger shaft e partly revolves, and in doing so arm g moves out, tightens straps y, i, pulls over picking arm \ and drives the shuttle. Spring and strap k are employed as in the lever pick to pull back the arm after the delivery of each blow. It will be observed that a loom furnished with this pick can be run in the opposite direction without dis- turbing a shuttle, the only effect produced being the slackening of straps i, j. As used on pick and pick looms, the alterations are : first, a double striker I at each side of the loom instead of a single one ; and secondly, a lateral motion, in addition to the rotary one previously mentioned, is given to the short shafts e by box tappets and lever connections, for the purpose of moving picking fingers d out of reach of strikers 1, When once moved, the shafts retain their positions for two picks. Before leaving the system of picking from the bottom shaft it may be desirable to mention a contrivance, which, although clumsy in construction and never looked upon with favour by the majority, is still used, and un- doubtedly possesses good points. Its chief advantages consist in the gradual development of power, and in maintaining a fixed energy in the picker for all speeds of loom. Two flat springs are bolted to the back framing at opposite sides of the loom; also at right angles to and above a cam upon the bottom shaft, they have their forward ends connected to the picking arms, and are in turn gradually elevated, but suddenly liberated at the proper moment to deliver a pick. 288 MECHANISM OF WE A VI NG PART Picking from the Crank Shaft The principle of picking from the crank shaft appears to have been introduced by Smith Brothers in 1834, with a view of utilising the superior speed and power of the fly wheel as compared with those of the bottom shaft, to obtain a more powerful pick and reduce the consumption of motive power. If in two looms, otherwise equal, the circles described by two strikers — one moved by the bottom, and the other by the top shaft — have the same diameters, then a blow delivered by the latter will be four times as powerful as that delivered by the former, because the speed of one is twice that of the other, and 1^ : 2^ : : 1 : 4. On the other hand, a pick delivered from the crank shaft must be com- paratively sharp and harsh, owing to the short time during which striker and picking finger are in contact; also to the impossibility of paying much attention to shaping either striker or receiver. On turning one of these looms slowly over and slightly pressing a picker against its box end, it becomes apparent that a picker is not under the control of the striker for more than \ the length of shuttle box. Whereas a similar experiment made with a cone pick shows the picker is pushed fully |- the length of shuttle box before contact ceases between picking tappet and cone. One application of this principle, known as the " carpet pick," is shown in front and side elevations, Figs. 163 and 164 : is a portion of the crank shaft, h one of the fly wheels, c a striker bolted upon the outside of An inclined shaft is placed outside the end framing and supported by footstep / and upper bearing g ; it carries PICKING 289 short finger collar A, and curved arm i; the latter is connected to picking arm h by passing strap ] round its shoe in the usual manner. Shaft e is elevated diagonally by a box tappet /, keyed upon the bottom shaft, that causes lever m to oscillate on pin by means of bowl 0, running on a stud in and in the groove of // a fork is also formed at the extremity of which bears against collar A, and raises or leaves in Fig. 163. Fig. 164. their normal positions lever m and shaft according to the shape of tappet I ; if the former, finger d is higher than striker c, therefore picking arm A remains stationary ; if the latter, d and c are brought together, and the pick is delivered. By altering the form of tappet I and driving it at a suitable speed, two or more shuttles can be driven from one side of the loom in succession in this, and also in the two following modifications : — Alterations made in the method of moving picking u 290 MECHANISM OF WE A VING TART finger d have formed the subject of several patents. Two only of these will be mentioned here — namely, Yates's, and the swivel pick; they are shown in Figs. 165 to 168, and in each case the same letter refers to a similar part in the carpet pick. Yates (Figs. 165 and 166), instead of lifting e bodily, caused its upper end to be pushed away from the framing, Fig. 165. Fig. 166. by making use of an inclined plane ^, as an upper bearing for e ; ^ is furnished with two flanges, between which the top end of e moves as in a groove and rests against the curved face of cam I ; the latter is set-screwed upon a light shaft driven by wheel gearing from the crank shaft, and as it revolves, e is pushed up, and out, until d is finger beyond the reach of striker c. In the swivel pick (Figs. 167 and 168) the upper bearing PICKING 291 g only permits of a rotary movement in consequently e is neither lifted as in the carpet, nor pushed out of the way as in Yates's pick. Instead of which, finger d is hinged to and rests upon lever jp. To miss a pick tappet I moves round, its full side elevates ^ and ^, until d is above striker c. Fig. 167. Fig. 168. Scroll Picks Two varieties of scroll motions have been introduced by Smith Brothers, the object of one being to lift striker c (Figs. 169 and 170) out of range of finger d by means of a stationary scroll j9, bolted to the end framing and bored to admit crank shaft a. Scroll consists of beads cast on one side of a disc in such a manner that grooves 1, 2 are formed to run into each other near the top at point j?', where a movable piece g is curved inward for the purpose 292 MECHANISM OF WE A VING PART of turning half -moon r out of one groove and into the other, as clearly shown in detached figure (171). Half -moon r travels in grooves 1, 2, by the rotation of fly wheel 5, to which it is connected by a stud s passed K Fig. 169. through a hole in striker bracket and a piece ^, bolted to l\ permits u to rotate with and slide upon the side of I), thus moving striker c in and out — in, when half- moon r is in groove 1, and out, when in groove 2. PICKING 293 In the latter position c is brought into contact with finger (i, and the pick is delivered. Fiu. 170. Figs. 172 and 173 show the parts of a revolving scroll; its object being to provide a means of lifting a swivelling finger d above a striker c, secured to fly wheel I, A three- Fig. 172. passes under and supports finger and the remaining arm carries stud s, which passes through the hole in half-moon 296 MECHANISM OF WE A VING PART r. It is obvious that finger d will be raised when r is in groove 1, and dropped to pick when it is in groove 2. If this contrivance possesses advantages over the swivel pick they are not very prominent. Half-moons are un- doubtedly troublesome; they are frequently broken, and the general wear and tear is considerable. Scrolls do not lend themselves to conversion into pick and pick motions. Over-picks Several over-picks that were commonly met with a few years ago are fast becoming obsolete, and it may be affirmed of the system in general that, with one exception, it is not making headway. Still that one exception probably embraces as many looms as are to be found using all other systems and varieties of picking motions taken together. The cone pick is the one referred to ; it forms part of most fast-running looms weaving light and medium fabrics ; also, of some narrow and wide looms weaving heavy fabrics, and the mechanism has been modified to suit pick and pick looms. It consists of a vertical shaft a (Fig. 174), which, when placed outside the loom framing, gives the most satisfactory results, for in that position the picking tappet can be moved near its own point of support, and greater steadiness and solidity are thus obtained. Shaft a is fitted to the loom framing by a cannon bearing near the top, and a footstep at the bottom; a merely serves as the fulcrum of a lever, of which one arm is a short stud passed through a slot in a, and fixed in position by screw and nut, to carry a loosely fitting, conical roller. The other arm c is of wood, and much longer than the former ; it is attached to a ring on the top of shaft a. PICKING 297 that has radiating teeth on its upper surface for a ring with similar teeth on the under side to fit into, and a grooved cap is bolted fast upon picking arm c to hold all From the forward end of c a leather strap together. Fig. 174. 1 1^ 6^ passes down to the picker, the latter being retained immediately over the shuttle box centre by a spindle on which it slides. Motion passes through the cone-shaped antifriction roller upon stud h from a tappet nose /, keyed upon the bottom 298 MECHANISM OF WEAVING PART shaft and inside the framing, that in striking h causes shaft a to turn partly round and move arm c and its picker towards the loom centre with sufficient velocity to drive a shuttle across. Tappet / is made in three parts, one a fixed disc, another is similarly shaped, but provided with a smooth, oblique surface, and concentric slots for bolts to pass through, the slots permit of a slight adjustment to vary the time of picking; a groove is also formed to take the third piece — namely, the picking nose. Mechanism is fitted at the other side of a loom to corre- spond, except that picking tappets / are diametrically opposed to each other; when the nose of one points up, that of the other points down. The conical roller on h is kept in constant contact with tappet /; and picking arm c and the picker are both moved back after the delivery of each pick by a spiral spring ^, which is hooked upon a convenient part of the framing, and into a strap that is fastened by a set screw to a hoop upon a, immediately below a bead of the framing. Or, what is better still, for narrow looms, the straps from both shafts a are connected to opposite ends of a spiral ^, situated below the warp, then as a pick is delivered at one side, the power required to distend g is utilised in pulling back the other arm c, A picking strap is always more or less slack when arm c is in a state of rest ; it therefore follows that a picker is only partially moved back by c, so it requires the shuttle to impinge against it to complete its backward traverse. The high speeds attained by many looms using this pick have rendered it necessary to bestow considerable attention upon its construction with a view to easy working, small consumption of power, and reduced wear and tear ; but it is still defective, and must remain so, for the energy X PICKING 299 is developed from a comparatively slow-running shaft in such an exceedingly short time that in ordinary looms the crank shaft only makes from \ to | of a revolution, there- fore at a speed of 200 picks a minute the shuttle must make its journey in from -g-^-^ to part of a minute ; but its velocity is not uniform ; it is obtained from a tappet that pushes the cone stud, in equal units of time, through evenly accelerating spaces in the proportion of 1, 3, 5, 7, 9. It is difficult to obtain with accuracy the amount of power required to drive a shuttle, owing to the variable conditions under which it performs its office, such as differences in friction, caused by large and small sheds, rough and smooth, close and open warp, strong and weak swells ; but it can be approximately reached by using the following rules : — First, work accumulated — ^ , where w equals the weight of shuttle in pounds, v velocity in feet per second, and g 32-2. For example, assume a loom to make 200 picks per minute; that a shuttle weighing 10 oz. is moved across a space of 5 feet in f of a pick, the average speed of 200 X 8 X 5 shuttle is ^ = 44*4 feet per second, and the energy 60 X 3 developed is 1 ^ ^ ^ ^ j - 19*13 foot lbs. per pick, or ^ 16 X 2 X 32-2 ^ 19-13x60 ^ ^ = ot a horse-power. Now, what must be the force of a blow to produce this velocity of 44*4 feet per second % Let the mean force of blow in lbs. = 19-2, and the distance in feet through which it acts = 10^ Then ^^'^ - 23*04 lbs. per pick. 300 MECHANISM OF WE A VING PART Or the question may be put thus — A body at rest is acted on by a force which gives it a velocity of 44*4 feet per second in t seconds. The body weighs 10 oz. Find the magnitude of force. Let F equal force in lbs. Now, v=ft .*. /=-; here is taken as = 44*4. To find the numerical value of assume a picking shaft to revolve 100 times per minute, and that it turns through an angle of 22J° to deliver each blow. 360° = tV of Pai't of a minute = yV t w ^ xf f o = 22-5 ^ of a second = ^. , , , .44*4 Knowing t we have f - — — -, which gives a numerical z 44*4 X 80 value for /, the acceleration = = 1184. o . , ^ w ^ 10 X 1184 .11 Also - —J- — — A6 lbs. per pick when energy stored up in swells is neglected. The length of picking arm c (Fig. 174) is to some extent determined by the width of loom, as the latter should not exceed c by more than 2-^ times (thus, on an average, a 20" arm is used with a 45" shuttle race). The length of picking arm fixes the position of upright shaft a, because, when reed and fabric are in contact, arm c, if drawn over the centre of picking spindle, should place the centres of picking strap (where it leaves the arm) and spindle in the same vertical plane. In order to prevent a waste of power by giving the picking force an upward or a downward direction, it should be transmitted to the cone stud h at an angle as nearly approaching a right angle as possible consistent with the intensity required, for much of the harshness of these picks results from inattention to this matter. PICKING 301 It is clear that as the cone moves horizontally in the arc of a circle, its surface will constantly form different angles with the tappet shaft, and therefore to the picking tappet /; hence the latter is peculiarly bevelled to present a flat surface to the cone in every position. This bevel has rendered the mathematically correct construction' of such tappets somewhat difficult. For many years it was a practice to fix a roughly shaped wooden model in^ position upon the tappet shaft and rotate both slowly to see if any sharp edges were presented to the cone, and if so, to alter its shape where necessary ; after which the model was used as a pattern to cast from. A picker may move 12'' or more at each stroke of a picking arm, but it is only driven through 8'' to 10'' by tappet /, the inertia completes its movement and is of no consequence to the shuttle, since the latter leaves the picker as soon as a decrease in speed takes place. A picking tappet may be said to have three functions. It must gradually tighten the picking strap to avoid giving a jerk to the shuttle ; during this time arm and therefore stud move through an angle approaching 10*^. This is followed by picking the shuttle across, and 22|° of accelerating movement in c, d is usually allowed for the purpose. Lastly, a further movement is given to c, in order to bring them gradually to a state of rest and thus reduce wear and tear — say 7|° in the decreasing ratio of 3, 2, 1. Some writers affirm that a shuttle is intended to be in the middle of the slay when the driving cranks are on their back centres, but in practice this is not adhered to. Much depends upon the time of picking, and a pick is set to act from 15° before the bottom centres are reached to 10° after they are passed. If, for example, movement in arm c begins when the cranks are on their bottom centres 302 MECHANISM OF WEA VING PART X and the tappet angles equal those given above, the cranks will be 10' + 221° + 71° = ^ 2 = 80° above that point, or 10° below the back centres when the pick is completed, and the shuttle will be at or near the slay centre when the cranks are on their back centres. Fig. 175 shows the construction of a picking tappet: a is the bottom shaft, h the picking tappet, c sectional view of the upright shaft, d cone, and e the horizontal plane in which d moves when nose / of tappet h is pressing upon it. Solid lines represent d in contact with the circular part of 5, and dotted lines show it when nose / has moved it to the extremity of its traverse — say 40°. As the tappet disc is employed to tighten the picking strap, and as the time taken to accomplish this is unimportant, we will begin with the actual delivery of the pick. Divide the 22|-° of move- ment in cone d into any number of parts — say 5, in the proportions of 1, 3, 5, 7, 9 ; then divide the remaining 7|° into 3 parts in the proportions of 3, 2, 1, and drop perpen- diculars down to cut line e. Continue the latter inside tappet 5, and describe a circle g from centre a to touch e. Draw a line radiating from a to pass through point 1 on make an angle of 30° and divide it into the same number of parts, but all equal, that the cone angle has been divided into — namely, 8. Through each point draw a line tan- gential to circle then with radia 1 to a, 9 describe arcs in rotation that cut the tangential lines. Let each point of intersection represent the centre of cone d as it is pushed outward by the tappet, and describe a circle round each having a diameter equal to that of cone d^ where it touches line li on tappet h. To find the centre line of nose /, trace a line that touches the periphery of each circle. The inner and outer edges of / can be obtained by following a similar rule, but the diameter of cone d must be taken Fig. 175. 304 MECHANISM OF WE A VI NG PART where it touches that portion of the tappet face to be drawn. It will doubtless be observed that the line already traced is slightly inaccurate, owing to the use of one uniform diameter of circle to represent the cone instead of making them all differ in accordance with the part of cone in con- tact with tappet and again, that only one true circle will be given, when cone stud and tappet shaft are parallel ; at all other places elliptical figures should be described. But the plan followed has been adopted to avoid undue complication, also because the inaccuracy is so exceedingly slight as to be of no practical account. Still, if it is thought desirable to construct a geometrically correct tappet, take the dimensions of cone d at the eight points shown, and construct an ellipse of the proper shape round each corresponding point where circles are shown. An idea is prevalent amongst weavers and others that a shuttle is less liable to fly out if a fall is given to a race board, between the shuttle box and centre of reed space, of about in a 45'' loom, and if the box spindle is higher, and farther from the box back at the inner than at the outer end. Some loom-makers also bow the reed to the extent of on the above-named width of loom. The reason assigned is that a picker in travelling up an inclined plane elevates the rear end and depresses the forward end of a shuttle, and by so doing reduces any tendency in the latter to rise from the race board. Also that a sloping race board and a curved reed assist in further reducing the chances of flying out by permitting the shuttle to continue moving in a downward and backward direction until the centre is reached, and by that time much of its energy has been expended. Advocates of this plan point to under-picked looms as X PICKING 305 demonstrating its utility, for in slot picks the circular movement of the picking arm, when near the front of a box, has a tendency to press down its picker, together with the rear end of shuttle, and thus impart an upward direction to the front of a shuttle. As most of these looms are admittedly more liable to throw shuttles out than over- picks, a certain amount of plausibility is seen in the con- tention. Still the difference may be solely due to the more rapidly developed force in an under-pick, and until reliable experiments have been made, the question must remain in the domains of pure theory. Positive Picking The system of positive picking preceded that of negative picking by many years ; it dates back to the days when attempts were first made to weave by automatic machinery, and has since been a favourite problem with many in- genious men of each succeeding generation. Neverthe- less, little real advancement has been made for ordinary weaving. The most extensively used motion of this class for other than small-ware looms was introduced by Lyall in 1876 ; it certainly contains many good points, and is less complicated than some negative picks, but whether through defective details or the staunch conservatism of Englishmen, it has met with little favour in this country ; still, it is steadily making headway in the United States. Lyall's mechanism is illustrated in Figs. 176 and 177 : a is a disc crank turned by a side shaft and bevel wheels from the driving shaft ; it gives motion by means of con- necting rod & to a sliding block in a slotted vibrating arm c. Link d is pivoted to the framing and attached to the X 3o6 MECHANISM OF WEA VING PART sliding block ; it causes the latter to move up and down the slot as crank a revolves. At its upper extremity c carries a pinion and also a band wheel 6, the former engages with the teeth of a curved rack /, which rotates e as arm c oscillates. A shuttle band g is wound round Fig. 176. the periphery of e, whence it is led over sheaves at opposite ends of slay % and finally connected to a carriage y, which moves upon four wheels on a bed in the slay bottom. The wheels run in pairs on short horizontal pins fixed Fig. 177. near the ends of carriage two of which are seen in Fig. 177, numbered 1, 2. Wheels 3, 4 project above and are journalled in so that they touch and are turned by con- tact with 1, 2. The shuttle Ic rests upon and is drawn along by carriage j. Shuttle wheels 5, 6 press against the insides of wheels 3, 4, and top wheels 7, 8 are only in PICKERS 307 touch with and roll along the under surface of bevelled rail /. This arrangement holds the shuttle down in its place and flying out or getting off the carriage becomes impossible. Assuming a right to left motion to be given to carriage and shuttle, the wheels revolve in the direction indicated by the arrows in Fig. 177, and all with the same surface velocity. Wheels 3, 4 lift the warp threads successively for wheels 5, 6 to roll over them, therefore a lateral movement of warp is precluded. The shuttle moves at a constantly accelerating pace to the centre of fabric, whence its velocity is correspond- ingly diminished until a pause is reached with the opposite side. This arises from two causes : the first is due to crank a and the connecting pin of sliding block, and the second to the action of link d. The shuttle begins to move when crank a is passing round its front or back centre, where for an instant the pin in the sliding block is in a state of rest ; but immediately one of those centres is passed the pin begins to move slowly, and increases in speed until at the bottom or top centre its forward velocity almost equals that of crank pin, then a corresponding decrease takes place up to the next centre. In the second place, as arm c moves from either extremity of rack /, the block is pushed by link d nearer the fulcrum / of arm c, and a greater velocity is developed by so doing. PART XI PICKERS The picker is more severely taxed than any other part of a loom, especially when it receives the shuttle, for the MECHANISM OF WE A VING PART Mow is given with a steel point and not distributed over an 'even surface. It has been estimated that a blow of an ordinary calico shuttle is equal to 1 lb. weight falling 3 feet ; and as these looms commonly run at 200 picks per minute, it is not to be wondered at that the wear and tear of pickers form a serious item of expense in a weaving mill. Pickers are made in a variety of forms and materials according to the requirements, or fancy, of the loom-maker and manufacturer. The principal materials used are buffalo -hide, leather, wood, iron, and brass, but the first named is in by far the most general use. To be of real service they should be of [good quahty and of equal weight. A variation of \ of an ^ounce should not be exceeded for one make of loom and fabric. Picker - making is a separate industry, and numerous processes are requisite before a finished article is obtained, which are briefly as follows : — Steeping the skins in water to soften them, cutting them into strips of dimensions to correspond with the size of picker to be made— ordinarily about 3|'' by T'O"; they are then rounded, notched, punched, passed through rollers and shaved down to a uniform thickness, matched, or the weight equalised by adding narrow strips which are to be coiled inside in the fol- lowing processes — namely, drilling, inserting a staple, riveting, roughly shaping, and pressing, by powerful machinery, into a usable shape ; they are afterwards slotted, punched, and drilled, in which condition they are sent to the manufacturer, but are by no means ready for use. When new they contain moisture and are more or less pulpy. They should be dried in a place where there is a con- •stant current of moderately warm air for six weeks or PICKERS 309 more, then steeped in good Gallipoli oil for at least one month, and are afterwards hung up and gradually dried for from six to eight weeks. The times named are often considerably exceeded with good results. If pickers are not thoroughly dry before being steeped in oil, the oil will not allow the moisture to evaporate. Strapping Good strapping is of equal importance with good pickers ; the best quality of leather, although necessitating a greater outlay at first, is cheapest in the end, for it does not stretch so much or break so soon as inferior qualities, and looms fitted with it are not stopped so frequently for repairs, consequently there is a double gain : first, in a greater output from a loom during the time a strap is in use ; and secondly, from the superior lasting powers of a good, compared with a bad article. Picking Bands A picking arm has a ring groove made near its forward end into which the picking band fits, but the methods of attachment are very varied, and great difference of opinion exists as to what is the best length of band, and which is the most economical way of connecting it to arm and picker. Whilst some prefer bands of from 12'' to 14'' long by to \\' wide, others contend that bands from 23" to 50" long possess many advantages over shorter ones. Both single and double bands are used, some being directly connected to arm and picker, others are connected to a secondary or even to two secondary straps. Leather prepared in the usual manner is employed by §ome MECHANISM OF WE A VING PART manufacturers, but on the majority of looms, green or oak tanned bands are found. The following are a few of the methods of applying bands to a cone pick on narrow looms : — One end of a tab 8" to 9'' long by \" to \\' wide is pushed through the side and top slots of a picker ; both ends are brought together, and a slit of about ^' is made near the ends for the picking band to pass through ; the latter, 23'' to 24'' long, is bent round the ring groove of the arm, sewn across and down each side, punched with holes |" to |" apart, inserted into the slotted piece, and secured by pressing a wooden peg, or leather plug, through a pair of the holes made to facilitate length adjustment. A method of using a double band that is often preferred to the foregoing requires a strap from 23'' to Tl" long, with a longitudinal incision made near its centre, and after fitting it in the ring groove, one end is drawn through the slit and pulled until both ends are level, then, as in the previous arrangement, both ends are passed through a short slotted tab on the picker and fastened with a peg. A piece of string tied round the band, immediately below the arm, holds it in position. Short single bands require a double tab, 9" long by If" wide, to be turned round the arm, where it is sewn, and riveted with a wire staple to keep it in the groove ; it is also slotted to receive the upper end of a band, 12" to 13" long, that is slit near one end, then passed through the front and top grooves in the picker, after which the other end is threaded through the slit, drawn tight, and made fast by a peg to the tab on the arm. Single bands, 2e5" to 50" long, have one hole punched near the lower end, which goes in at the front and comes out at the picker top; a piece of leather pushed PICKERS into the hole prevents the band slipping out when working. The free end is provided with an indefinite number of holes approximately f apart ; it is taken up to the arm and twisted round spirally until all is wound on, then a wire pin, driven into the arm, goes through one of the holes to retain it in position. A wire staple or a strip of leather screwed to the arm bridges the gap of the ring groove and holds the band in its place. Instead of a leather plug to connect band and picker, it is by no means unusual to find bands slit near the picker, and the other end passed through the slit to couple them. Those in the habit of using long bands contend that the consumption is smaller than where short ones are employed, for the reason that bands generally break near a picker, and in the case of a long one the damaged part can be cut away, a fresh hole punched, a portion of the band unwound from the arm and again secured at both ends as before. The Check Strap is a simple and effective contrivance for assisting swells in preventing shuttles rebounding. It was invented by Crook and Eccles in 1845, and is now in a slightly modified form attached to quick-running looms, with either fast or loose reeds. It is merely a leather strap a (Fig. 178), about \" wide and of equal length with the loom. Four guides h are screwed to the slay front to support and permit of a sliding motion in a, which at its centre has a thick piece of leather c, Y l^^g ^7 ^' deep, nailed upon it between the two centre guides, that are about apart, to allow 3J'' of traverse. Each end of a is secured to a short tab ^' long and wide ; both tabs have a hole punched 312 MECHANISM OF WEAVING PART near one end to allow the picking spindle to pass freely through, and a slot is made at the other end wide enough to receive strap a. When in position a is fastened by a piece of wire, or instead of slotting the tabs, a buckle may be sewn upon each, for either provides a ready means of adjustment. Tabs are placed behind pickers, and all should be put upon the spindles at the same time, but strap a is not adjusted until the loom is ready for work ; it is then usual to place the cranks upon the bottom centres and tighten a until the picker is about Y' from a box end. Two pieces of leather e called guards, broad by ^" long, are punched at both ends, then one end is pushed on a Fig. 178. spindle in front of a tab, and the other goes behind a spindle plate. If e are properly adjusted piece c can be removed and the check strap will work satisfactorily. In no case must a picker be allowed to strike a box end, or shuttles will rebound, cops be knocked off, bobbin or shuttle pin broken, and it will be apparent that the pick is too strong, the swells too weak, or the check strap is imperfectly adjusted. Buffers Buffers are required to keep a picker from striking a spindle stud ; they consist of three or four thicknesses of leather, 6'' long by \" wide, doubled to bring all the ends together, and then riveted with a wire staple, but loops are XII SWIVEL- WE A VING left large enough for the spindles to pass through. A slot is provided near their extremities, each to receive a strap, one end of which is bolted to the slay front, the other is pushed into a buffer slot and fastened by a piece of straight wire. Other buffers consist of a strap doubled at one extremity, and punched to receive the spindle, the free end is screwed to the slay front. Others, again, are formed of a closely coiled strap, with a hole drilled through all the coils for the spindle to enter, after which the straight end is made secure upon the slay. PART XII SWIVEL-WEAVING Swivel-weaving stands in the same relation to picking that lappets stand to shedding. Both are intended to pro- duce small continuous stripe figures, with material differing from that composing other parts of the fabric, or small detached figures may be diapered over the surface of the piece. Both give somewhat the appearance of embroidery, and afford a means of preventing waste, for the reason that figuring material is only used where figures are to be made. If the two systems are compared as to beauty of effect, variety of detail, and general excellence of workmanship, swivels are vastly superior to lappets ; but when compared for cost of production, the latter is found to be more economical ; for swivel looms are costly and elaborate ; they are run at a slow speed, and each pick of swivel weft is driven above a pick of ground weft ; hence the length of MECHANISM OF WE A VING PART fabric produced is determined solely by the number of ground picks inserted. Swivels have been made in power-looms for upwards of twenty years, but they are still, to a large extent, pro- duced on hand-looms. Owing to the large addition to the number of shuttles in use, the increase of mechanism neces- sary to manipulate them, and the absence of detectors to stop a loom if figuring threads break, the weaver's work is considerably increased, and frequent stoppages of the loom are entailed. Swivel-weaving consists in adding ribbon shuttles to an ordinary loom in such a manner that they can be held out of the way, dropped upon the race board, and moved under lifted warp at pleasure, but there must always be a space between spot and spot at least equal to the length of a shuttle which varies from \" to 5". There is, however, an attach- ment on most looms for moving all the shuttles laterally after the first line of spots has been completed, so they shall occupy positions exactly midway between their former situations, then the second line of spots prevents unsightly rows, and gives an equal distribution of figure. Or they may be thus moved to form twill and other arrangements. In power-looms, swivel shuttles are fitted in a movable carrying frame attached to the front of a slay, which is raised and lowered by the indirect action of a Jacquard, operating through a cam, levers, rods, and spiral springs. The frame can thus remain stationary for any length of time by simply leaving cards blank where they face the needles that govern it, and by suitably perforating cards the frame can be lowered in unison with the formation of a shed. The Jacquard is also employed to throw the ordinary picking motion out of action whenever swivels are to be moved, and to cause long racks to drive the small shuttles. XII SWIVEL- WE A VING 315 All cams used should be driven by a clutch, or be capable of sliding on a key-way, so that they can be put in and out of gear by the shedding motion, for many fabrics have their spots arranged at such distances apart that certain portions only contain ground weft ; hence, during the weaving of the last-named parts, the frame is lifted, and all its connections are inoperative. A shuttle is hollowed out to take a weft bobbin, which is loosely fitted upon a movable wire spindle, one end of the latter is pushed into a hole inside the shuttle, the other end is bent at right angles, and pressed tightly into a slot in the opposite edge. A spring presser held against the bobbin centre prevents the weft unwinding too freely ; it is assisted in maintaining an even tension by two glass rings situated on opposite sides of the shuttle eye, each being secured to a thin spiral spring fixed in the shuttle ; the springs have a tendency to take up slack weft, and so prevent loops forming round the edges of the figures. After threading weft through both rings, it is led through an eye at the shuttle centre. A more perfect tension results from the employment of a spindle, in which a hole is drilled to receive one end of a thin spiral that has a collar fitted to its opposite end. A sleeve, with four protruding flat springs, goes over the spiral, and fits tightly upon the collar. The weft bobbin is, in turn, lightly pressed by the flat springs. The spiral, however, gives this method its great advantage, for as weft is drawn away there is a tendency to uncoil the spring, but immediately weft becomes slack, the spring exerts sufficient force to wind back four to five inches. Shuttles are sometimes supported in their frame hori- zontally, sometimes vertically. If they are horizontal, a single row is employed, and portions of the frame between 3i6 MECHANISM OF WE A VING PART the shuttles are cut away to leave room for lifted warp when the rack is on the race board. The back of each shuttle is grooved throughout its entire length, and a lip is formed to fit into a similarly shaped holder near the bottom of the frame. A rack, with metal or leather teeth, equal in length to and secured upon the top of every shuttle, gears with two pinions. Motion is given to all pinions by a vibrating rack, long enough to reach across the reed space. If the long rack moves to the right, all shuttles will be driven into holders immediately on the left, and by reversing the direction of motion they will be restored to their former positions. It is therefore obvious that holders must exceed shuttles in number by one. A rack is usually governed by the Jacquard through a cam, a series of links, and an upright shaft, and means are provided for putting the rack out of action when- ever it becomes necessary to stop swivelling. Figuring shuttles occasionally move simultaneously with the ground one. In such cases it becomes necessary to open two sheds — the lower one for ground weft, and the upper one for figuring material ; but it is questionable whether the gain is a real or an imaginary one, for the great strain put upon a few threads lifted for figure will undoubtedly cause numerous breakages ; if, on the other hand, the inventor's aim could be realised, the production of a loom, with a single line of shuttles, would be doubled (see Howarth and Pearson's Jacquard, Fig. 92). When shuttles are supported vertically, four or five may be placed one behind the other, and, as a consequence, four or five colours may be employed in the development of each figure. All shuttles are suspended from frames, in which a separate groove is cut to receive every line ; the latter are also grooved and flanged,, and racks are secured XII SWIVEL- WE A VING 317 to their tops which engage with pairs of pinions pivoted in a fixed holding frame. Shuttles placed in this manner must become shallower as they recede from the reed in proportion to the slope of the warp, but they are somewhat difficult to adjust and refill. In order to render the last- named processes easy of attainment, each holder is hinged to the bottom edge of a fixed frame, therefore any set of shuttles can be instantly turned up, and thus brought into a handy position for manipulation. Circles Circles are used as substitutes for swivel shuttles in hand and power looms ; they can be successfully employed where not more than two lines are required, but beyond that they are less useful. Compared with the older method, a greater number can be placed in a given width ; still this advantage is purchased at the price of increased parts. A circle frame is raised and lowered in a similar manner to a swivel frame. The chief differences are to be found in the circles which consist of metal plates shaped to resemble horse-shoes, each is secured by, and turns inside, two flanged plates, but, in a state of rest, all points are towards the warp. Every plate carries on its face a weft bobbin, and at the rear a series of pins project far enough to gear with the teeth of a long rack capable of moving laterally. When a frame is lowered, the openings in the shoes permit spotting warp to rise, then movement is given to the rack, and all the circles make a complete revolution. In doing which, the weft bobbins pass under the warp at one side and emerge at the other. On the following figuring pick rota- tion is reversed. 3i8 MECHANISM OF WE A VING PART PART XIII WARP-PROTECTOES, OR FAST AND LOOSE REED MOTIONS If, from any cause whatever, a shuttle fails to reach its proper box, the loom must either be instantly stopped by a fast reed motion, or provision made whereby a shuttle may remain amongst the warp, and the cranks revolve with- out doing damage. In this case, a loose reed motion is used. All looms are provided with curved levers called swells, which, in fast reed motions, serve the twofold purpose of protecting warp from being broken when a shuttle is in the shed, and also of stopping a shuttle from rebounding after entering a box. Swells a (Figs. 179 and 180) are generally made of iron, but many wooden ones are still to be met with, although the latter rapidly wear away at the shoulder, and it becomes necessary to replace them with new ones, or to cut away portions from each end in order to again bring them back to their required shape. This long formed one of the chief defects of wooden swells, but a simple and satisfactory method of preventing wear, and also of increasing or diminishing at will the obliquity of their surfaces, has been recently introduced by Messrs. Collier, Evans, and Riley of Farnworth. The invention is shown in Fig. 180, where swell front a has a coating h of thin steel firmly secured upon it at c, and at a screw is passed through a slot in h to hold the rear end in position. At ^ a hole is bored through the wood to receive a set screw /, that is retained in position xin WARP- PR OTEC TORS 319 by a plate g affixed on the swell back and a lock-nut. The end of / projects beyond the wood and carries a second nut li for the purpose of impinging against the Fig. 179. inside of plate J, hence by screwing or unscrewing / the size of shoulder can be regulated ; h also provides an elastic cushion for the shuttle to strike a2:ainst. A 320 MECHANISM OF WE A VING PART swell is hinged near the outer end of a shuttle box (except in a few looms where the pin is reversed), and its curved face protrudes inside box h (Fig. 179), and is pushed forward by a lever c resting behind it. Below h an oscillating rod d works in bearings, preferably fixed upon the swords ; but, in the sketch, is shown attached to the slay bottom ; it extends across the loom, and has levers c set-screwed upon, and flattened blades e welded to it. In front of, and directly in the path of a buffer /, known as a frog, is provided with a shoulder that is struck by blades e whenever a shuttle is absent from a box; / rests upon a portion of the framing, and its forward end presses against a rubber cushion ^, a spiral, or a flat spring, and Fig. ISO. in many looms frog / is connected to a brake on the fly wheel by an arm, in such a manner that a forward move- ment of / puts a drag upon the wheel, and assists to stop the loom speedily. Blades e and frogs / are shaped to dovetail into each other when in contact — this is done to avoid slipping ; and in that position, without other aid, they could bring the loom to a sudden stand, but breakages would be of con- stant occurrence, through the violent shocks of impact between e and / when the driving strap was exerting its full force to carry the slay forward. This defect is over- come by attaching an arm A to one of the frogs / and shaping it to place its forward end immediately behind setting-on handle h when the loom is in motion; so, as the frog recedes from the blow of blade arm A partakes XIII WARr-PROTECTORS 321 of a similar movement, pushes handle h out of its notch, and the driving strap is thrown upon the loose pulley. Either a flat spring I presses against finger c to assist in holding blade e in striking position, or the top of a spiral spring hooked upon the under side of e has its lower end fastened to a bracket bolted upon the slay. So long as a shuttle is moving properly the loom con- tinues in motion ; for, as the former enters a box, it pushes back swell a and finger c, the latter causes rod d to oscillate and carry the front of e clear of /, but if a shuttle does not reach a box there is nothing to push back swell conse- quently the loom is stopped. Beyond the primary condition that all parts must be true and free to move as intended, the points requiring most attention are : length, position, and pull upon the blades, together with the size and shape of swells. The blades, or blade, — for some looms have a stop motion at one end only, although two are better able to keep the slay from twisting, ^ — should be long enough to leave a shuttle in the w^arp without fear of breaking threads as the slay moves forward to beat up ; but if the blades are too long, there is a tendency for them to strike the frogs before a shuttle has time to lift them clear away ; they ought to rise about Y above the frogs when a shuttle is full in its box. As previously stated, a rod shoidd not be fixed to the slay bottom, but upon the swords, and its best position is found if blades are horizontal when in contact with frogs, for then the rod centre sustains the shock, and there is no tendency to twist ; but if a rod centre is above the frogs, its tendency to rotate is proportionate to its elevation, and part of the energy of impact tends to bend or break finger c, Y 322 MECHANISM OF WEAVING TART If springs upon the blades are too strong, a shuttle will require driving with increased force, and worn shuttles and swells result, together with a waste of motive power. Worn swells, narrow shuttles, and wide boxes are trouble- some ; each and all cause a loom to knock off when it should continue in motion, just as worn blades, frogs, and weak springs allow a loom to run when it should be stationary. At the best a fast reed motion is harsh in action and puts great strain upon many parts of a loom, but the rigidity of the reed especially recommends it for the pro- duction of heavy fabrics. Loose Reed Looms A contrivance, patented by Hornby and Kenworthy in 1834, greatly improved by BuUough in 1842, and also by many later inventors, remedied most of the evils attending the use of the stop rod, and is extensively used on light and medium looms. By its aid a loom can be stopped if a shuttle gets trapped in the warp by simply pushing the setting-on handle out of its notch and leaving the loom running until the driving strap is entirely upon the loose pulley. The parts of a loose reed motion are shown in Fig. 181, where a is the reed, with its upper rib pushed into a groove in slay cap the groove forming a sort of pivot for a to swing upon, c is a strip of wood, a bar of flat or of angle iron extending across the loom and pressed against the bottom rib of a by the action of a series of arms, one of which is marked d ; they support c, and are fixed upon an oscillating rod 6, occupying a similar position and cor- responding with the stop rod of a fast reed loom. An additional arm with an adjusting screw is also secured to XIII WARP-PROTECTORS 323 rod and its screw presses against the rear of a slay sword to regulate the pressure upon reed a. Two curved levers / are keyed upon e equidistant from opposite ends of the slay ; these, as the slay moves forward, have their outer ends carried against the upper or lower faces of Fio. 181. wedge-shaped pieces g called heaters or frogs ; the latter are held rigidly by brackets h bolted to the framing. An arm i is also keyed upon rod e exactly opposite the working point of starting handle j ; but when reed a is in its normal position the point of i must be below the corrugated surface of a buffer ^ upon 7. Another arm I is fixed 324 MECHANISM OF WE A VING TART upon rod e outside the end framing, which carries an anti- friction roller at its extremity, so that, as the slay falls back, / is taken into contact with a bent spring m bolted to the framing. The function of parts m is to hold reed a steady as a shuttle moves across, and spiral springs n perform the same office when m are apart ; n is stretched upon two hooks, one on rod e and the other on the slay sword. The reed is held firm to beat up, but its lower portion must move back freely if a shuttle is trapped. Fingers / should be underneath, and from f ' to ^ P^^t the point of frogs g when reed and fabric are in contact ; at the same time bar c should bear tightly upon reed a, to prevent any vibration and give a smart blow to the weft. If the cranks are on the top centres, only the power of springs n keeps the reed from falling back ; there should then be a space of about 2'' between fingers / and frogs ^, so, in case a shuttle stops in the shed, the warp resistance will be sufficient to open springs 7^, force back reed a and bars thus causing rod e to oscillate, fingers / to be carried to the upper slope of frog cj, and arm i to abut against buffer h It is thus seen that fingers and frogs combine to remove the weight of bar c and the pull of springs n from the reed at the instant that arm i pushes the setting-on handle out of its notch, and the loom is brought to a stand without violent concussion. But a loose reed is not suitable to the pro- duction of heavy fabrics, owing to the difficulties experi- enced in making the reed sufficiently firm to beat up, and sensitive enough to throw out the reed before a shuttle can damage the warp. Swells in a loose reed motion prevent a shuttle from rebounding out of its box, and to some extent assist in XIII WARP-PROTECTORS 325 guiding it ; but if they are defective, cops are liable to be knocked off, and shuttles, pickers, and bands rapidly wear out. Many of them are similar to that described as forming part of a fast reed motion, but the fulcrum pin has a tend- FiG. 182. ency to act as a wedge and leave the greater part of a swell out of contact with the shuttle when in a box. Messrs. Harker and Grayson have patented a double spring swell which controls the shuttle better than most, for it constantly changes its position with relation to the shuttle, and yet, as the latter is driven home, the entire length of swell acts upon it. Fig. 183. Figs. 182 and 183 are plan views of this swell : a, a is the back board, c bolt and screw respectively, used for fixing spring d in position ; at e the spring projects slightly in front of a, but at / it stands about yV' beyond. When a shuttle is received or released it comes into contact with the spring at and d gives way ; but when it 326 MECHANISM OF WE A VING PART is more than half-way in its box, the spring gives way at / and acts in conjunction with d, PART XIV SHUTTLE-GUARDS Some simple trustworthy appliance that will prevent shuttles from flying out without entailing more labour upon the weaver is still a desideratum, notwithstanding the fact that a large variety of apparatus is now before the manufacturing public professedly for attaining the objects named above. Few guards will absolutely prevent a shuttle from flying out, but most of them prevent flying up ; and as the latter evil is more to be feared than the former, it must be granted that to this extent they are serviceable ; and if more of them were constructed to withstand the severe shocks occasioned by beating up, it is probable the number in use would be largely increased, but so long as light flimsy brackets, and screws driven into the slay cap are common, the people most concerned in keeping them in order will use every endeavour to prevent their application, for after a few months', or even weeks' work, the brackets break or become loose and require constant attention. A guard to be really serviceable should be secured to the slay by bolts that pass entirely through the cap, and where parts move in bearings the latter should be broad and firm. Guards are of three classes — ^ namely, rigid, semi- automatic, and automatic. For heavy work the first-named appear to meet with most favour. They consist of at least two metal plates, ^' to 5" long by 2" broad, both of which XIV SHUTTLE- GUARDS 327 are bolted through cap and sword ; these plates or brackets support opposite ends of a fixed rod yV' to ¥ in diameter that extends over the shuttle's course at a height of from '2Y to o' above the race board, and from 2'' to 1Y in front of the reed. If the loom is a wide one, intermediate brackets are bolted upon the cap to serve as stays. Such a guard is firm and simple, but it seriously interferes with the weaver's freedom to see and repair broken warp. It cannot prevent a shuttle from leaving the reed and flying out at the edges of a fabric, for there is no guide of any kind at these points, unless an additional iron plate is bolted on the slay front near the entrance to each shuttle box. Such plates are made to bend inwards and over to the cap ; they are also spoon-shaped, to ensure catching a flying shuttle, but what a weaver gains in safety he loses in having his labour increased when changing shuttles. Semi-automatic guards are numerous, and varied as to detail. Hall & Sons make use of two brackets which they attach to slay cap and swords, as previously described, but they are used as bearings for a swivelling guard bar, made of round iron, to move in. When the bar is in working position it is locked by forcing it against fixed stops, and to the rear of two tapering pins situated in chambered re- cesses in the brackets ; a spiral spring is pressed into each recess, and the shank of a tapered pin goes inside the spring, through a hole at the back of the chamber, and is secured by a nut. Each pin will therefore slide back when suflicient pressure is brought to bear upon it, and this is accomplished by grasping the guard bar and drawing it forward before entering warp into the reed. When not in working position the bar rests upon the slay cap, and does not in any way interfere with the weaver ; but before re- starting the loom he must turn it down again. The only 328 MECHANISM OF WE A VING part increase in labour consists in moving the bar up and down by hand, beyond this the contrivance is simple and firm. The guard patented by Marriott can be more truly de- scribed as semi-automatic than the preceding, for it will turn down and lock itself in position when a loom is set in motion. The guard blade a (Fig. 184) is approximately Y broad by Y thick ; it is above the shuttle race and 3'' from reed to front of guard. Two brackets h are secured to the slay cap by bolts that enter at the reed groove and pass out at the top of the cajj. Blade a is fixed to h by hinges c ; that nearest to the starting handle is made with a cam-shaped projection, behind which a sliding wedge- shaped piece cl drops to lock blade a. Wedge d is slotted and fastened to bracket & by a bolt. When the loom is stopped to repair warp, d is lifted by the forefinger, and the guard plate is turned up to the cap by the thumb. On Marriott's Guard A Fig. 184. XIV SHUTTLE' G UA RDS 329 re-starting the loom, a falls down as the reed strikes the cloth and wedge d locks it in position. This is also a simple inexpensive motion, and is not liable to become deranged through vibration. Hamblet k Clifton Hamblet & Clifton introduced a guard in which two brackets screwed to the slay cap form bearings for a light shaft ; one has a clutch face, and is chambered out to admit a helical spring ; one end of the latter is secured to the bearing, and the other to a collar on the shaft. The collar has also a clutch face, and is pulled against the similarly shaped face of the chambered bracket by the spring. The shaft carries three arms curved downwards to support two parallel rods. In case threads are to be drawn in the rods are turned up by hand, and the inclined surfaces of bracket and collar open the spring. Vibration caused by beating up is sufficient to ensure the guard assuming its proper position when the loom is re-started. Automatic Guards The third class of shuttle guards is also a large one, and must in all cases contain parts that automatically move out of the way when a loom is stopped, and are automatically restored to position as soon as a loom is re- started. The following has not been selected because of its superior merit as compared with others belonging to the same class, but simply as illustrating the general lines on which inventors have worked. It was brought before the notice of manufacturers in 1893 by Farrell, and consists of 330 MECHANISM OF WEA VING PART a bowl a (Figs. 185 and 186) that turns freely upon a stud in the starting handle. A curved lever h rests upon a in such 4t Fi«. 185. H uiaiiiicr that as the handle is pushed outward h is elevated. A stout wire c connects b with the outer arm of a lever d, XIV SHUTTLE-GUARDS 331 retained in bearing and situated on the bottom rail of the end framing. A similar wire / connects the inner arm of d with the cranked end of a rod g that extends en- tirely across but behind the cap ; g is supported near each end by brack- ets li screwed upon the slay back, and at intervals between those points by bearings also attached to the cap. At two places — namely, outside the reed ' ends and inside the swords — four holders i pass through slots in A and protrude over the shuttle race, two on either side of the loom, but one above and resting upon the other. A hole is drilled entirely through the headpiece of each holder for the purpose of supporting two parallel guard rods j. The holders are connected in pairs by pins to brackets U set-screwed upon rod g. When the loom is put iji motion bowl a acts through levers 5, and connections to oscillate, and thrust forward holders i and rods } over Fio. 18G. j\ to cause rod g 332 MECHANISM OF WE A VI NG PART the shuttle's path ; but whilst the upper rod moves out from the reed, the lower rod attains a position fully an inch in advance of its fellow. This effect is produced by making slots of in the upper pair of holders and simply drilling holes about behind the guard rod of the lower pair ; then passing two pins in each bracket h through holders one entering the slot of the top piece, and the other the hole of the bottom piece. Immediately a loom stops the pins draw back holders and rods until the latter rest against the upper part of the reed face, in which position they offer little obstruction when warp is being drawn in ; but the motion has the de- fect, common in shuttle -guards, of lacking rigidity and strength. PART XV BEATING UP Beating up follows picking, and is the function of the slay. To successfully control its movement a swift blow must be delivered to drive weft home, and movement must be retarded to afford time for a shuttle to pass across. During the infancy of the power-loom inventors were somewhat at a loss to discover a good method of obtaining such a movement. Flat springs placed behind the swords were tried, the slay was pulled back slowly by the action of cams, and when the shuttle had passed across, the swords were suddenly liberated, and the stored-up energy in the springs was employed for beating up. An adjust- able screw rendered it possible to obtain a light or heavy blow. XV BEATING UP 333 In the pneumatic loom the slay was pulled back along grooves in the framing by two weights, each hooked upon a separate strap ; the latter, after passing over guide pulleys, were made fast to the slay back. Two cylinders having the usual piston and rod connections w^ere em- ployed, and the rods w^ere also attached to the slay. At the proper moment compressed air was admitted into the rear end of each cylinder to drive the whole forward with the requisite velocity. Both these appliances permitted a dwell of any duration, and the intensity of the blow was unaffected by the velocity of other parts of the loom. To this extent they were w^ell adapted for the work to be done, but against such advan- tages must be set the disadvantage of moving a slay negatively, and thus depending largely for a constant stroke upon the regularity with which the fabric was drawn forward, for it is obvious that any variation in this respect would increase or diminish the forward traverse of the slay. After the power -loom had been brought into more general use mechanism was introduced which embodied the principle of placing the slay under positive and con- stant control. Of the advantage of which there can be no question ; but a result of doubtful utility followed — namely, that of altering the force of impact between cloth and reed with the speed of loom. Several positive appliances have been used, such as para- bolic wheels, cams, and cranks. When the reed was required to give a sharp stroke upon the cloth, or when it was necessary to give as much time as possible for the shuttle to pass. Dr. Cartwright made use of an elliptical wheel, which he fixed upon the main driving shaft, and geared it with an eccentric secured 334 MECHANISM OF WE A VING PART upon a crank shaft, to cause cranks to move the slay in such a manner that a large side of the elliptical wheel geared with the small side of the eccentric to give a superior speed for beating up, and a small side of the elliptical wheel geared with the large side of the eccentric to move the slay at an inferior speed, for, one was equivalent to a large driver and a small driven wheel, and the other to a small driver and a large driven wheel. Any such arrangement is, however, superfluous except for very wide looms. When it is remembered that all machinists and manu- facturers were familiar with the use and capacity of the cam, it is perhaps scarcely to be w^ondered at that this piece of mechanism appeared to them suitable for govern- ing the slay. One application consisted of two short shafts placed at opposite ends of the loom, to each of which a pair of fast and loose driving pulleys were attached. At the inner ends of such shafts a grooved cam, shaped to give the requisite eccentric movement, was fixed. Each connecting arm carried an antifriction roller that worked in a cam groove, and as the cams revolved, the slay was alternately pulled and pushed at varying speeds. The difficulties arising from the use of two driving straps, and the consequent tendency for the slay to get slightly askew, together with other contrivances equally unmechanical, prevented this method from ever coming into general use, still, cams are used at the present time. Probably the best application has a main driving shaft ex- tending from end to end of the loom, and the cams are placed upon it at each extremity, thus obviating the neces- sity for more than one driving strap, and at the same time keeping the slay parallel with the fabric. XV BEATING UP 335 Cams can be readily shaped to impart any degree of irregularity to a slay. In modern looms the slay consists of a rocking-shaft a (Fig. 187), supported in cannons bolted to the end framing. Near each extremity of shaft a two upright arms or swords C B Fig. 187. h are fastened; they are capable of adjustment, and support the slay sole c, — a heavy piece of wood strong enough to resist the constantly recurring shocks given as weft is driven home ; it extends across the loom, and upon its upper surface a thin race board is fastened to serve as a guide for the shuttle. A longitudinal groove e is made 336 MECHANISM OF WE A VING PART near the back of c wide enough to receive the lower rib of reed /. The slay cap g is similarly grooved on its under side to take ,the upper rib of /, and when bolted upon swords h the reed is firmly fixed in position. In single shuttle looms the shuttle box bottom forms a continuation of but if two or more shuttles are employed separate boxes are fitted at one or both ends of the slay bottom. Two cranks li are bent on the main driving shaft, to which connecting arms h are fastened by metal straps, gibs, and cotters. Arms h are also fulcrumed to swords h by connecting pins I. The revolving cranks set up a swing- ing motion in the slay, and when the latter is pulled from the fabric a shuttle is driven across ; as it is pushed towards the fabric weft is forced into position by the reed. When cranks were introduced for beating up it was known that a variable motion could be obtained from them, but for some time it was not clear as to how the parts could be accurately adjusted to give a required amount of variation in velocity between that for beating up and that for the passage of a shuttle. Engineers, however, over- came that difficulty, and now any one who wishes to become thoroughly conversant with the crank can do so by study- ing a good work on the steam-engine. In order to apply cranks in the best manner, it is essen- tial that the points where their usefulness may be effected shall be minutely examined. These points are : (1) Eelative positions of cranks and connecting pins ; (2) diameter of circles described by cranks ; (3) length of connecting arms. Eccentricity is modified by altering the height of connecting pins with relation to the crank shaft centre ; BEATING UP 337 therefore it is desirable that the best positions of both shall be defined. If the beam steam-engine is taken as a guide, and similar parts of both machines are placed in the same relative positions, they will then be capable of imparting a steady and useful motion to a slay. Let the rocking-shaft of a loom be represented by beam centre a (Fig. 188), and a slay sword by half beam then crank c, connecting rod and connecting pin ^, are similar in both. When the beam of an engine is horizontal, the centre z 338 MECHANISM OF WEAVING PART of connecting pin is in the same vertical plane as the crank shaft centre ; hence, if the beam oscillates, the pin e will move through equal angular spaces above and below its position in the horizontal beam ; and there will be an equal and minimum divergence from the vertical plane. In order to bring these parts of an engine into the positions they must occupy in a loom, it will be necessary to move them on centre a through an angle approaching 90°, as shown in Fig. 189 ; let point a represent the centre of rocking-shaft, and vertical line h a slay sword, for in the majority of looms the swords are vertical when reed and fabric are in contact ; c the centre of connecting pin is, in narrow looms, frequently not more than Y^' behind the vertical line, but in some wide looms it is carried back to the extent of 14'' by casting ears behind the swords; this is done to permit of short connecting arms and to avoid unduly enlarging the cranks. With a, c as radius describe arc cZ, to show the path of c when motion is imparted to it ; next, find the position of c at the end of its journey by marking from c on line d point c equal to the diameter of crank circle to be used. Bisect line c, c', and draw line a, a', which will pass through the centre of pin c at c\ exactly midway between the extreme points of its oscillation. A line, 6, drawn at right angles to a, a , and passing through c\ will contain the centre of crank /. To fix the position of /, mark upon line ^, with c as centre, the length of con- necting arm, and the relative positions of the parts com- mon to loom and engine are maintained, and will permit of any degree of eccentricity being given to a slay. This is, however, a question affecting loom-makers in designing- looms more than manufacturers when using them. Valuable eccentricity in a slay is represented by the difference in its velocity when the cranks are passing BEATING UP 339 through a given number of degrees above the front centres, and when they are passing through the same number of degrees below the back centres. If the rotary motion of a crank is converted into re- ciprocating rectilinear motion, it is found that equal angular spaces in the former produce unequal rectilinear s^^aces in the latter. With a long connecting arm the difference is slight, but with a short one it becomes very marked, as will be seen in Fig. 190, where a is the crank circle, 340 MECHANISM OF WE A VI NG PART and 1 to 8 are equal angular spaces, h the length of con- necting arm, and c a point to be moved in a horizontal plane. If the crank is moved To the end of division 1 , point c will have travelled from c to 1 At the end ,, 2 „ „ 1 5, 2 and in like manner back again to 8. It will be observed that all the spaces 1 to 4 are unequal, 8 7^-"- O- ••S^ C I 6, 5 7,/<^^'' 81 3 4 5 8 Fig. 190. 2 being the largest and 4 the smallest, but the difference between spaces 8 and 4 is of most importance, as 8 repre- sents the speed of slay for beating up, and 4 its speed for a shuttle to pass across. Cranks and connecting arms cause the slay to move irregularly, because the arms are pulled in a more or less diagonal direction, which carries one end nearer the crank centre and holds the other end at a fixed distance from it ; therefore a crank in moving through 90° from front to bottom centres will pull the slay through a space equal to the radius of crank circle, plus the difference in length XV BEATING UP 341 between a straight and a tilted connecting arm when measured on a horizontal line; further, as the crank passes from the bottom to the back centre, the movement of the slay equals the radius of crank circle, minus the in- crease in length of arm due to changing it from a tilted to a straight position, if measured as stated above. By means of a graphic sketch the precise amount of variation can be found in a few minutes. Fig. 191 is an example of this : the circle described by crank, &, c length of connecting arm. When the crank has passed from c to arm h\ c is tilted, and therefore shortened if Fia. 191. iC measured on line b. To find how much it is diminished in length take b, c as radius, and from point b' describe arc c, 3 ; space 2, 3 represents the loss in length, and c, 3 the actual movement of connecting pin. Two things are obvious : first, that the crank has caused the pin to be moved through space c, 2 ; and secondly, tilting arm c has further moved it through space 2, 3 ; hence, in solving this problem, these are the two points to be determined. Calculations can also be made to ascertain the variation ; thus, if the crank circle has a radius of 2V, and the con- necting arm is 10" long, the actual movement of connecting pin between, a top and front centres, and b bottom to back 342 MECHANISM OF WE A VI NG PART centres, can be found by considering that two sides of a right-angled triangle are given to find the length of the third, for radius of crank circle equals altitude, length of connecting arm equals hypothenuse, and base is required. The length of base is found by squaring both sides, sub- tracting one square from the other, and extracting the square root from the remainder = \/ 10'^ - 2*5^ = 9 '68'' = the length of tilted armor a loss of 10 - 9-68 = 0'32'' ; hence actual movement of connecting pin for each quarter of the crank's revolution is Top centre to front . 2-5" radius of circle + 0-32 = 2-82'' Front centre to bottom . 2-5^' „ + 0-32 = 2-82" Bottom centre to back . 2-5" „ -0-32 = 2T8'' Back centre to top . 2-5" „ - 0*32 = 2-18" 10-00" the complete movement of connecting pin for one revolu- tion of crank. It is more difficult to determine how far a connecting pin travels for other than 90° or 180°, owing to the want of sufficient data — namely, the altitude of the triangle formed ; but by the aid of trigonometry, or by the use of the following table of natural sines and versed sines, the movement is readily obtainable : — XV BEATING UP 343 Table showing the Motion of the Crank (Crank = 1 inch) Degree. Sine. Versed Sine. Degree. Sine. Versed Sine. Degree. Sine. Versed Sine. 1 •0174 •0001 31 •5150 •1428 61 •8746 •5152 2 •0349 •0006 32 •5299 •1519 62 •8829 •5305 3 •0523 •0013 33 •5446 •1613 63 •8910 •5460 4 •0697 •0024 34 •5592 •1709 64 •8987 •5616 5 •0871 •0038 35 •5735 •1808 65 •9063 •5773 6 •1045 •0054 36 •5877 •1909 66 •9135 •5932 7 •1218 •0074 37 •6018 •2013 67 •9205 •6092 " 8 '1391 •0097 38 •6156 •2119 68 •9271 •6254 9 •1564 •0123 39 •6293 •2228 69 •9335 •6416 10 •1736 •0152 40 •6427 •2339 70 •9397 •6579 11 •1908 •0183 41 •6560 •2453 71 •9455 •6744 12 •2079 •0218 42 •6691 •2568 72 •9510 •6909 13 •2249 •0256 43 •6819 •2686 73 •9563 •7076 14 •2419 •0297 44 •6946 •2806 74 •9612 •7243 15 •2588 •0340 45 •7071 •2929 75 •9659 •7411 16 •2756 •0387 46 •7193 •3053 76 •9702 •7580 17 •2923 •0437 47 •7313 •3180 77 •9743 •7750 18 •3090 •0489 48 •7431 •3308 78 •9781 •7920 1 Q ■if / 041/ .o/i on 7Q / y y o i o • 0009 20 •3420 •0603 50 •7660 •3572 80 •9848 •8263 21 •3583 •0664 51 •7771 •3706 81 •9876 •8435 22 •3746 •0728 52 •7880 •3843 82 •9902 •8608 23 •3907 •0795 53 •7986 •3981 83 •9925 •8781 24 •4067 •0864 54 •8090 •4122 84 •9945 •8954 25 •4226 •0937 55 •8191 •4264 85 •9961 •9128 26 •4383 •1012 56 •8290 •4408 86 •9975 •9302 27 •4539 •1089 57 •8386 •4553 87 •9986 •9476 28 •4694 •1170 58 •8480 •4700 88 •9994 •9651 29 •4848 •1253 59 •8571 •4849 89 •9998 •9825 30 •5000 •1339 60 •8660 •5000 90 1^0000 1^0000 The sine of any degree is the length of a line dropped from a point in the periphery of a circle at right angles to the diameter line. The versed sine is the space between the sine and 344 MECHANISM OF WE A VING PART periphery of circle when measured on the diameter line. For example, let h (Fig. 191) be the diameter line, c' a point 60° above a, then c\ d is the sine and also the altitude of triangle h'\ d, c'\ h"\ and d^ c is the versed sine which shows the movement of pin due to crank. If the circle has a radius of i'\ and the arm is \T long, the sine of 60° being 0 '866025, and the versed sine 0'5 for a circle 1" in radius, then for a circle in radius 0*866 X 4 = 3*464'', and 0-5x4 = 2'' .-. Vl2^-3-462 = ll-5" the length of base, and also inclined arm, or a loss of 12" - 11 -5" = 0*5", and 2"+0*5" = 2'5" the motion of connecting pin. The eccentricity of a slay can be obtained by taking a given number of degrees at both front and back centres, ascertaining the movement at each place, and subtracting one result from the other. It will be noticed that the more a connecting arm is tilted the greater the eccentricity ; hence it follows that increasing the throw of crank, shortening the connecting arm, or both, will increase eccentricity, just as diminishing the throw of crank and lengthening the connecting arm will produce opposite results. Consequently if it is desired to obtain a definitely in- creased or diminished amount of eccentricity, the rule already given is applicable. For example, a 2|" crank in turning through an angle of 180° from the front centre, will, during the first 90° of movement, pull the connecting pin of a 10" arm 2*5 + 0-32 = 2*82", and for the second 90° it will move 2*5 - 0*32 = 2*18", hence 2*82 - 2*18 = 0*64" eccentricity. Now what size of crank must be used with a 10" arm to give 0*8" eccentricity? 0-64 : 0-8 : : ^'b- : 7-8125 x/7-8125 2-795" the crank required. XV BEATING UP 345 For Proof. x/IO'^- 2-7952 = 9-6014 10 - 9-6014 = 0-3986'' due to tilting the arm. 2-795 + 0-3986 = 3-1936 2-795 - 0-3986 = 2-3964 •7972 eccentricity. To determine the length of stroke to be given to a slay, width of loom, make of fabric, and to some extent the breadth of shuttle, must be taken into account. Width of loom, because revolutions of cranks must be diminished as width is increased ; now this obviously results in decreasing the energy stored up in a slay, whilst energy required is in proportion to increased width of fabric ; therefore to maintain energy, weight of slay and diameter of crank circle should be increased and the con- necting arms proportionately shortened. Although it is well known there is a speed above and below which a loom does not increase its pro- ductiveness, manufacturers do not follow any definite rule in fixing the relative velocities of looms differing in width but producing similar fabrics ; nevertheless, this is a subject that each might determine for him- self by comparing the lengths of weft consumed in a given time by looms weaving wide and narrow pieces. Then, if the loom using most weft is taken as a model, the best plan for speeding one of any other width is to cause the shuttles in both to move at the same velocity. For instance, if one slay 50'' long from box to box, weighs 150 lbs., is moved by 3|-" cranks and 14" arms, and another slay 100" long is moved by 4|" cranks and 9 ' arms, both producing similar fabrics, and it is desired to 346 MECHANISM OF WE A VING PART find what relation the revohitions and weight of the latter should bear to those of the former. The revolutions will be in inverse proportion to width .*. 100 : 50 : : 2 : 1, but force of impact must be in proportion to width .'. 50 : 100 : : 1 : 2, if both move at the same velocity ; but this is not the case, for one has 3 J'' and the other 4^'' cranks, so to find the velocity of each, points in the circles 20° above the front centres are taken. The sine of 20°- 0-34202, and versed sine = 0*06037. 0-34202 X 3-5 = ri97" 0-34202 X 4-5- 1-539" 0-06037 X 3-5 = 0-211" 0-06037 X 4-5 = 0-272" x/l4^- 1-197^ = 13-94" 14 - 13-94 = 0'06" the shortening of the 14" arm. x/ 92 - 1-5392 = 8-86" 9 - 8-86 = 0-14" the shortening of the 9" arm. Therefore, as the versed sine of the 3 J" crank is 0-211", the connecting pins will have moved when the cranks are on the front centres 0'21 1 + 0-06 = 0-271". The cranks move their connecting pins 0-272 + 0-14=0-412''. Force varies as the square of the velocities .*. 0-412^ : 0-271^ : : 150 lbs. : 65 lbs. If the cranks in two looms of equal width made the same number of revolutions per minute, a slay weighing 65 lbs. would deliver a blow^ equal in weight to that of 150 lbs., assuming the former to be governed by large cranks and short arms, and the latter by small cranks and long arms ; but it has been shown that force required is as 1 : 2 .*. l'-^ : 2^ : : 65 : 260 lbs. the weight of slay. Velocity at any other angle smaller or larger than 20° can be obtained in the same manner. A reed moves through a space greater than the throw of XV BEATING UP 347 crank, because the connecting pins are placed below the point of its contact with the fabric, and the difference is in proportion to the distance between those points ; thus, if centre of rocking-shaft is 28'' from centre of connecting pin, and Z\" from fabric, and a crank describing a circle of 5'' is used, the movement is as 28 : 31 : : 5 : 5*33". Looms constructed for weaving light and medium goods have the swords placed vertical when reed and fabric are in contact to prevent passing and repassing the centre of rocking-shaft at each revolution of the cranks, and some claim for this position that the blow is firmer because it is delivered at right angles to the cloth, but this is clearly not the case, for cloth is invariably depressed between breast beam and harness ; hence to strike a blow at right angles the swords must lean towards the cranks when the latter are on the front centres. In wide and heavy looms other considerations have weight, probably the principal one being that their large cranks would carry the race board too low when the slay was pulled back, to permit of the bottom warp line touching it, and unless this contact is maintained, shuttles are in constant danger of flying out. To prevent which it is not unusual to find the swords vertical when the cranks are on the top centres. But whether the swords are vertical, inclined forward or back- ward, the rule previously given for finding the relative positions of rocking - shaft, connecting pin, and crank centres can be used, provided the first line erected repre- sents the most forward position of the reed. One thing must be insisted upon — namely, that whatever position the swords occupy when reed and cloth touch, the race board must be bevelled until it is parallel with the bottom shed, as the shuttle is moving across. In a few looms the slay swings from above, but as this 348 MECHANISM OF WE A VI NG PART system is fast becoming obsolete, it is not necessary to devote any space to an explanation of the scheme of working. In carpet and certain other heavy looms it has been found necessary to give the weft two blows in quick suc- cession, and the plan adopted is to actuate a knuckle joint by a crank in such a manner that it is straightened twice for each revolution of the crank, and in straightening the weft receives a blow. In Figs. 192 and 193 a is a sword, 4 Fig. 103. h a crank, c a connecting arm, d the knuckle joint, e is hinged to a part of the framing, and / is attached to the slay by the usual connecting pin. In one sketch the knuckle joint is straight, but when crank h moves from point g to the top centre, it is pushed up, and when c reaches point li on the opposite side, the joint is again straightened. The time that elapses between the first and second straightening can be altered by moving the crank centre up Fig. 192. XVI THE REED 349 or down ; in the former it will increase, and in the latter it will decrease. PART XVI THE EEED is used to beat up weft after being inserted by the shuttle ; it helps to keep warp threads in their proper positions ; it forms a back guide for the shuttle to run against ; and it determines the fineness of the fabric. Prior to 1738 reeds were made of split cane, or reed; they were split by pressing reeds against a taper spindle, from which knives radiated at the required distance apart ; but in the above-named year, John Kay used flattened brass or iron wire instead, and the substitution was every- where welcomed as an important improvement that left the reed almost perfect. Eeed- makers call the flattened pieces of wire dents (namely, teeth), and they are made from the best Swedish iron wire, which is flattened to any extent by passing it through rollers ; it is next straightened on its edges by taking it between two plates and forcing it in a serpentine course amongst pegs ; the ribbon is then cut to a uniform width equal to that of the dent, rounded on the edges, smoothed by files placed to act on every part of the wire as it passes forward, straightened on the flat sides, and sent to the reed- maker, who passes it through a machine that automatically cuts it into equal lengths, inserts each in position, winds, presses, and entirely finishes the reed at the rate of from 300 to 400 dents per minute. A finished reed consists of a series of parallel flat wires a (Fig. 194), secured at their extremities by wooden 350 MECHANISM OF WE A VI NG PART rods approximately half-round in section, and bound together by a tarred cotton cord c, passing between the dents and round two rods. The cord usually consists of fourfold 32^ twist, folded to suit the space between dent and dent ; but round iron wire is occasionally used in place of cotton. For cotton fabrics reeds have from 6 to 90 dents to the Fig." 194. inch. They are sometimes made double by placing two sets of dents between ordinary laths, so that one set comes opposite the spaces of the other. Such reeds are used to keep the loose fibres of unsized warp from twisting and choking the shed. A reed is so constructed that a definite number of dents will be contained on a given length ; this is termed the count, pitch, or number of the reed; but the naming of reeds has been complicated without giving the slightest XVI THE REED 351 advantage to anybody. This is proved by the large number that have been discarded within the last thirty years in favour of two or three more rational methods. The following classified list of some reeds still in use will probably place the matter in its simplest form : — Eeeds named from the number of dents contained on one inch — Eadcliffe. Huddersfield. American. Keeds named from the number of dents contained on a certain number of inches — Stockport, 2". Scotch, 37". Macclesfield, 36". Eeeds named from the number of groups of dents on a certain number of inches, each group consisting of 20 dents and called a beer, or porter — Bolton, 241". Bradford, 36". Dundee, 37". Worsted, 54". On 19 dents to a group, or beer — Dewsbury, 90". On 5 dents to a group, or beer — Holmfirth, 12". In some silk reeds, pitch, ends per dent, and width go together as 2000 reed, 8 thread 24", or 1800 reed, 4 thread 18". Of the foregoing the Stockport reed is in by far the most extensive use, and it bids fair to supersede most, if not all, others in Lancashire. The most useful comparison of reeds is by the number of dents per inch for a number 1 reed of difi'erent counts. To find which use the following Rule : — counts X dents per beer inches in basis Tims Stockport, = 0 -5". Bolton, , = 0 '8247". ^ ' 2" 24-25" 352 MECHANISM OF WE A VI NG PART Kadcliffe, No. 1 reed — 1 LlcliL Ubl incli. TT n n n PT"SiTi pi ^ i ?j — ^ 55 55 '5 — ^ 55 55 5J — 0-^ — u o „ 55 55 — \} \J Si i \J Zt 55 Macclesfield 55 — 0*09 77ft 55 Bolton 55 = 0*8247 55 Bradford 57 0*5556 55 Dundee 55 = 0-5405 55 Worsted 55 = 0*3704 55 Dewsbury 55 = 0*2112 55 Holmfirth 55 = 0*4167 55 To convert the count of one reed to its equivalent count in another reed — given count x dents per beer x inches in basis of required inches in basis of given x dents per beer of required Example. — What does a 40 reed Bolton count equal in the Dundee and Stockport systems 1 40 X 20 X 37" 24*25" X 20 40 X 20 X 2" 24-25" X 0 = 61*03 Dundee reed, and = 66 Stockport reed. PART XVII MOTIONS TO STOP THE LOOM IF WEFT IS ABSENT When automatic looms were introduced it at once became apparent that additional parts were necessary to ensure their practical utility. Dr. Cartwright saw that some appliance must be added to stop a loom immediately weft gave out, so that a fabric should not be rendered unsightly by the presence of great cracks. XVII WEFT STOP MOTIONS 353 To meet this difficulty he fixed a wire inside the shuttle and near its eye to support a swinging staple, the latter to be held in a horizontal position by the unbroken weft ; but when a fracture occurred the staple assumed a ver- tical position, and was caught by a hook placed inside a groove in the slay and near the entrance to one shuttle box. This hook acted through a lever upon the driving belt, and brought the loom to a stand. Dr. Cartwright's plan was not satisfactory, nor was any striking success achieved by later inventors until the intro- duction of the "fork and grid," which was dated 1831, and claimed by Gilroy; but it was patented in England by Eamsbottom and Holt in 1834, improved by Kenworthy and Bullough in 1841, and left in 1842 by James Bullough in much the same condition in which it is found to-day — namely, with a brake attachment to make its action certain and instantaneous. From the last-named year this beautiful, delicate, and simple motion has held its position against all rivals, and now, except for special work, is applied to all looms. It is situated at one side of a loom between the fabric and shuttle box (see Figs. 195 and 196) : a resembles an ordinary three-pronged fork, having the prongs bent almost at right angles to the main part ; the end representing the handle is bent in the same direction about | of an inch from the end. It is freely balanced on a cross pin which passes through the horizontal part of the fork, but leaves the prong end somewhat lighter than the other. The fork holder h is at right angles to, and fixed by a set screw to fork lever c. The latter occupies a horizontal position immediately behind the setting-on handle c?, and is hinged at its outer end. The short bend of the fork rests upon the upper end of 2 A 354 MECHANISM OF WE A VI NG PART hammer lever which at that point is more or less semi- circular in form, and furnished with a catch correspond- ing with, and capable of taking hold of, the hook upon fork a. Hammer lever e is bell-cranked, and its lower end rests upon a cam /, fixed to the bottom shaft of the loom. As / rotates a vibrating motion is given to 6, causing its upper arm to move forward once at every second pick. Fig. 195. Between the reed and shuttle box a grid g is fixed vertically, so that the fork prongs can pass through without obstruction at each forward stroke of the slay. Its action is as follows : — When the slay moves forward weft Ifi is pushed against grid g by the fork prongs, and it ofi'ers sufficient resistance to prevent the fork from passing through, and also to cause the shank of the latter to tilt up at the moment when the hammer on e begins to XVII WEFT STOP MOTIONS 355 vibrate, consequently contact between them becomes im- possible ; but should weft be absent the fork prongs pass through the grid, and fork hook and catch remain in contact, the hammer moves forward and carries with it fork, fork holder, and fork lever. The latter pushes the setting-on handle out of retaining notch and its spring Fig. 196. forces the driving strap upon the loose pulley by means of lever and strap fork I. As this mechanism is only placed at one side of a loom it can only act on alternate picks of weft, but if the parts are properly adjusted the cranks will not revolve more than twice after weft is broken before the loom is brought to a stand. Considerable care is, however, required to so adjust the parts that a loom will stop when weft is absent and continue running when it is present. To ensure the 356 MECHANISM OF WEA VING part former, fork prongs must not touch any part of the grid or race board, and accurate setting is provided for by an adjusting screw m on the fork lever, which allows the prongs to be elevated or depressed, and the fork to be moved laterally to the right or left. A second setting screw n on the fork holder permits of a forward or back- ward adjustment at pleasure. If the fork passes too far through the grid bars there is a tendency to cut the weft, but if it does not pass far enough through, its hook will not be lifted above the hammer. Also if a shuttle rebounds weft is slackened, and will not then tilt the fork sufficiently. Any one of these things is sufficient to bring a loom to a stand when it should be running ; therefore, in order to keep a loom at work for its maximum time, fork prongs should pass just far enough through the grid to cause weft to lift the fork hook clear of the hammer catch as the latter begins its forward move- ment. A hammer lever, if accurately timed with the slay, will commence to move when reed and fabric are in contact, so that if weft is broken hammer e will catch the fork hook and throw the belt upon the loose pulley, but if weft is unbroken the fork hook will be kept up until e has moved forward far enough to miss it, and thus the loom will continue to run. The fork lever also requires attention; its movement should be just sufficient to push the starting handle out of its notch. The traverse of this lever is regulated by moving its fulcrum pin nearer to or farther from the fabric. Lateral movement in the slay will frequently cause the fork lever to tilt when weft is broken, and permit the loom to continue running. XVII WEFT STOP MOTIONS 357 The Brake has largely contributed to render the present high speeds possible, for previous to its introduction a speed of from 100 to 120 picks per minute was considered high. As ordinarily made it is simply a lever, a (Figs. 197 and Fig. 198. Fig. 197. 198), having one straight arm, upon which an adjustable weight & is hung, and the other arm, called the brake clog, is curved and covered with leather on the inside for the purpose of increasing its holding power. It is placed imme- diately below fly wheel c, and so governed that both can be instantly brought into contact, either when the weft breaks or when the loom is stopped by other means. 358 MECHANISM OF WE A VING PART The weighted end of a is connected by link d to a tumbler lever which rests upon and is in turn actuated by stud and bowl, or by a curved bracket / affixed to the setting-on handle. When the loom is started, e is raised by the pres- sure of / against its full side, and link by elevating the heavy end of a, moves the brake clog out of contact with c ; but if the weft fork acts e falls, and brake a bears upon fly wheel c, and brings the loom to a stand. It is questionable whether the ideal brake has yet been invented, for, simple as its function appears to a casual observer, it is attended with difficulties that have proved troublesome to overcome. A brake should not impart a shock or strain to any part of a loom when brought into operation. It should not reduce the impetus of a slay when the last pick is driven home, but should allow a loom to run freely until the driving strap has been traversed upon the loose pulley, and then bring it to a stand in the following half pick. It should always stop a loom at the back centre when the shuttle is at the fork side if weft is broken, but when stopped by the weaver for piecing broken threads the top centre is the most convenient place. It should not be any encumbrance or add to the labour of the weaver. In check and other looms, where it is essential that the pattern shall be unbroken, means should be pro- vided for holding the brake off when once removed until the proper starting-place, or pick, is found. An ordinary brake puts considerable strain upon the moving parts by acting before the driving strap is off the fast pulley, and thus, as it takes about one pick to move the strap, the crank shaft is forced against the top cap, the momentum of the slay is checked, the last pick is not properly beaten up, and the brake, by constantly acting in the same spot, wears the shaft, bearings, wheels, and XVII WEFT STOP MOTIONS 359 brake. The last named results in the loom stopping in different places. Some brakes serve merely as an appendage to the weft fork ; others act whenever the spring handle is moved. The latter, all things considered, appear to be preferable. Fig. 199. Haythornethwaite's brake seems to be constructed on original lines, and contains several good points. It is a double brake. The lever rt(Figs. 199 and 200) is actuated as in an ordinary brake by a finger I on handle % that pushes up the curved arm m of lever the last named has a connection & that supports the back brake clog c, and a second clog also attached to «, brakes at the front of 36o MECHANISM OF WEA VING PART fly wheel e. A spiral surface / forms part of but a groove g is left up the centre, except at one point where it runs diagonally to the outside. When the loom is working clog c is out of contact, and the neb at its extremity is in line with groove g : but if the weft breaks the fork lever begins to be drawn back as the crank is leaving the front centre, and pushes the spring handle off as the bottom centre is reached. Lever a immediately drops, and the neb of c enters groove ^, but the brake remains inoperative until the front centre is again reached, and with it the diagonal point of junction between /, then the neb of c is acted upon by Fk;. 200. the spiral surface of brake wheel /, which draws it upon its enlarging periphery, and the loom is brought to a stop every time in exactly the same spot, with the slay a little beyond the back centre, and the shuttle at the fork side. As clog c is pushed back it pulls clog d into contact with fly wheel 6, and slightly elevates the forward end of a until it rests against the wedge-shaped fixing li on spring handle when further upward movement in a is checked, and both faces of the fly wheel are held as in a clip with- out putting unnecessary strain upon the working parts or checking the momentum of the slay in beating up the last pick. Spiral k holds a steady when the loom is workinc:. XVII WEFT STOP MOTIONS 361 Friction pulleys have been advocated for loom-driving instead of fast and loose ones. It is argued that by their instantaneous action less power would be required to brake and pick a loom than at present ; the former because the driving power ceases at once with the severed connection, the latter because full power would be exerted to deliver the first pick. Centre Weft Fork An ordinary side weft fork is not well adapted to pick and pick motions, for two shuttles may be driven in succes- sion to the fork side of a loom, and the weft of one be absent without the loom being brought to a stand. As the number of shuttles is increased the difficulties become greater, so with a view to overcoming them a centre fork was intro- duced, and its inventor claims for it that the weft stop motion for pick and pick looms is perfected, also that it is available for alternate picking. As its name implies, this fork works in the centre of the reed space ; it is made to feel for each separate pick, and in case one is absent the loom is stopped. Figs. 201, 202, and 203 are plan, front and side eleva- tions respectively : a is the slay bottom in which a trans- verse groove h is cut to allow the prongs of fork c to enter and sink below the race board. A bracket screwed upon the front of slay sole, supports and guides the moving parts. Fork c consists of two prongs, hook and arm/ all fastened to and vibrating with pin cj. The fork is made to swing on the centre cj by the fork shifter A, which, as the slay falls back, rises and pushes up c far enough to allow a shuttle to pass beneath its prongs ; arm / regulates to some extent the swing of c by preventing it from moving too far 362 MECHANISM OF WE A VING PART in case a jerk results from the contact of h with e. This is done by / entering a slot in li and rising to the top, where further traverse is impossible. Two thin iron plates d! screwed upon d form grooves for shifter li to move in. Eod i is hinged to A, and passes 3 Fig. 201. through two holes, one drilled in the upper, and the other in the lower flanges of a forked guide Z:, which is free to swing upon holder / ; the latter is bolted upon the top of bracket m, A foot on m rests upon the floor, and the whole is secured by set screws to a cylindrical bar extend- ing across the loom. Eod % is furnished with an adjustable XVI 1 WEFT STOP MOTIONS 363 stop hoop y. It constantly presses against the upper flange of and must be set, when the slay is thrown back, to elevate fork c sufficiently. Guide Iv being immediately below shifter A when fork c is at its highest point, it follows that the space between the top flange of Z: and fork c increases as the slay moves forward, and therefore c, /i, i fall by their own weight until the prongs of c are arrested by weft w lying across gap h. The fork is prevented from falling- farther until its extreme ends are pulled by the slay beyond the weft. It then sinks to the bottom of groove ^, but meanwhile the slot in h has fallen too low for hook e to take into it, so the loom con- tinues in motion. On the other hand, if weft is absent, prongs c fall at the same speed as shifter A, and hook e sus- pends the latter from the end of its slot. A raised piece 0 is thus held opposite to, and the slay moves it into contact with, a swivelling finger ^, which turns on centre ^, and is fixed by a bracket against the inside of breast beam s. On the under side of s a lever ^, fulcrumed at its centre, Fig. 202. 364 MECHANISM OF WE A VING part has one end pushed back by finger r, and the other pushed forward and })ressed against arm 'c ; v being set-screwed upon shaft x', an oscillating movement is set up, and starting XVIII MECHANISM FOR GOVERNING WARP 365 handle y is moved in the usual manner to shift the strap upon the loose pulley. A centre fork is good in principle, but it requires careful setting, and if warp and weft are thin^ the weft is liable to be looped at the under side where the prongs act. PART XVIII MECHANISM FOE GOVERNING WARP Power-loom weavers have inherited the difficulties of warp management from their predecessors of the hand-loom, and the problem has baffled the inventive powers of machin- ists and others for fully a century, during which time hun- dreds of patents have been obtained, and probably thousands of experiments have been made without marked success. The causes of failure will be more apparent if the prob- lem is stated as it presents itself to a weaver. In the first place, a let-ofF motion should so act upon the warp that an equal strain will be maintained whether a shed is open or closed, and irrespective of the length of warp upon a beam, and without the necessity of adjustment from the beginning to the end of a warp. In the second place, warp should be positively delivered in lengths exactly corresponding with those drawn away by the taking-up roller. In the third place, the parts added should not be of such a complicated nature that more overlookers will be required to keep a given number of looms running. Little more than a century ago each warp beam had two holes bored through it near one end, and at right angles to each other. Into one of these, a rod h (Fig. 204), long enough to rest against cross rail c, was pushed, to hold 366 MECHANISM OF WEAVING PART the warp rigid during the weaving of from 2' to 3" of fabric, after which it became necessary to remove the rod from hole 1, unwind warp from the beam, and place it in hole 2. The defect of this system consisted in putting a con- tinually increasing strain upon the warp during weaving so long as h occupied a fixed position. About the year 1788 a weaver named Clarke adopted a simple contrivance that largely increased the production of D Fig. 204. a loom, and at the same time governed the warp more effectively. He coiled two ropes two or three times round opposite ends of beam a (Fig. 205), fastened heavy weights h to the outer, and light balance weights c to the inner ends of the ropes ; this, in addition to enabling a w^eaver to draw warp forward without going behind his loom, provided a vastly superior method of controlling it ; for to a large extent ropes and weights reciprocate with the shedding motion, but not to the extent of keeping an equal tension XVIII MECHANISM FOR GOVERNING WARP 367 upon the warp in all its varying positions ; yet, when a shed opens, warp is drawn from the beam, and when it closes, as much of it as is not used 11 in the fabric is Avoand on again. D Fig. 205. A further defect occurs at every revolution of the beam, produced by the extraction of a layer of yarn from its periphery, without simultaneously altering the circumfer- ence of the roller where the ropes act. To rectify which would necessitate weight being removed in proportion to the decreased diameter of coiled warp. 368 MECHANISM OF WE A VI NG PART For example, let the radius of a (Fig. 205) equal. 2|-", that of d ^\ and weight h 100 lbs. Then if warp is drawn from d until its radius becomes b'\ a weight of 6 : 5 : : 100 : 83*3 lbs. must be used to maintain an equal tension. Clarke's appliance slightly modified is found on y^y- of the cotton looms at present working in this country. One of the principal alterations has been adopted to prevent weights, applied in the manner described above, from settling upon the floor when used in a machine subjected to the shocks and vibrations that are common in a power-loom. It consists in removing balance weights, and fastening the inner ends of ropes, or chains, upon hooks in the framing, if stout fabrics are required, or upon flat springs similarly situated, for light goods, and their outer ends to weighted levers, either simple or compound \ but this renders a beam less sensitive to movement in the shedding harness, and less capable of taking back any excess of warp pulled from the beam than where balance weights are used. There are many points that require careful attention to cause a letting-off" motion to work satisfactorily ; of these the beam is of great importance. It is sometimes made of cast or wrought iron, but more frequently of wood ; of the latter there are two varieties, viz. solid and built ; in both dry sound timber not liable to warp should be used. An iron hoop is inserted into each end of a solid beam to prevent splitting, when bearded gudgeons, about Y' square in section, are driven in. Beam and gudgeons are then turned to a uniform diameter ; the former varying from b" to 6'', and the latter from ^' to \\' , A built beam is considered superior to a solid one, because its gudgeon passes entirely through the beam, and warping is conse- quently a rare occurrence. XVIII MECHANISM FOR GOVERNING WARP 369 Cast-iron ruffles are required to prevent ropes and chains from channelling the beam and thus rendering it useless ; they are occasionally turned to provide a smooth surface, but are often employed as they leave the moulding sand ; in any event, they must be smooth enough to allow ropes or chains to slip regidarly. Iron or steel flanges are fixed upon all beams intended for use in power-looms, to permit warps of great length to be wound on without fear of the outside threads slipping or becoming entangled. Each warp beam is supported in brackets bolted to the loom framing ; it must be parallel with back rest, breast beam, and taking-up roller, and the warp wound upon it with an even tension, or faulty cloth will be made. No definite rule can be given for weighting, nor for the length of warp exposed between beam and harness ; both vary with circumstances ; still, it may be taken as a general rule that warp should be held as tight as its strength will allow without giving the fabric a harsh feel or breaking the threads, else if weight is insufficient the fabri'c is apt to have a raw appearance. Too much or too little warp may be subjected to the pulling action of healds ; in either case un- necessary wear and tear results ; in the former, because it is pulled too frequently before being woven into the piece ; in the latter, because too great a tax is put upon its elasticity. From 20'' to 24'' is enough for certain looms, but 36" to 40" is not an uncommon length in heavy looms. Chains possess advantages over ropes for v/eighting many cotton warps in respect to more regular tension, and also as to the cost of keeping them in a fit working state. The latter is an important consideration to a manufacturer, for ropes cost from Is. to 6s. or even 8s. per annum per loom. The amount of friction produced by coiling ropes round 2 B 370 MECHANISM OF WE A VI NG PART beam ruffles depends largely upon their condition. New ones possess far greater holding powers than old ones ; and when made from different fibrous materials, they exert a varying frictional drag ; the condition of a ruffle also affects the tension put upon a warp ; but if a given rope is coiled round a beam, and has weight and counter attached to opposite ends, by doubling the weight of counter twice as much weight will be needed at the heavy end to maintain an equal pull; or, in other words, "friction is proportional to load/' A great difference will, however, be observed if an alter- ation is made in the number of coils of rope upon a ruffle ; thus if a rope is twisted once round a beam and weighted, a 1 lb. counter will be slowly pulled up by a 3 lb. weight on the other end ; if the rope is lapped twice round, nearly 9 lbs. will be required to elevate the counter, with three laps about 27 lbs. will be needed, and with four laps approxi- mately 81 lbs. will be necessary to perform the work. This is equal to the first, second, third, and fourth powers of the coils, or l = 3i, 2 = 32, 3 = 3^, 4 = 3^. Perry, in his Practical Mechanics^ gives the following results of actual tests with a 1 lb. counter-weight : — Laps of Rope Weight in lbs. required to slowly on Ruffle. raise the Counter. 1 1-6 s 4 2-1 1 3-0 li 4-0 5-1 n 6-6 2 8-0 Ol 10-0 24 14-0 21 20-0 3 23-0 3i 30-0 XVIII MECHANISM FOR GOVERNING WARP 371 The table shows a divergence from theory, but it is sufficiently accurate to demonstrate the utility of the rule given above. There is a point beyond which additional laps will not produce a beneficial effect in a loom, and it is reached when further slipping becomes impossible, for the weights are then wound up and the counters sink to the floor. With one end of a rope fixed to the framing winding up of weights must cause slack rope to be given off at the inside, and to some extent this induces slipping, but of so irregular a nature as to be absolutely useless to a weaver. Handy and simple as the rope and lever motion is, it is essentially unmechanical, and can only be considered as a makeshift for a let-off. When tested by the conditions stated on p. 365, it is found to be incapable of maintain- ing an equal tension, where sheds are opened and closed, or at different parts of the warp. An attempt is, however, made to approximate to a uniform tension during shedding by employing a vibrator which moves in to slacken warp as a shed opens, and out to tighten it as a shed closes, but the maintenance of a fixed pull from one end of a warp to the other is left entirely with the weaver who reduces the weight as the diameter of the warp beam diminishes. The second condition is not met in any way, for instead of warp being delivered positively, it is simply pulled from the beam as required by the shedding, beating-up, and taking-up motions. Still, faulty as this part of the contrivance undoubtedly is, here is its strongest point, for warp can be paid off and taken back as a shed opens and closes. The third requirement is admirably met, because nothing capable of doing the work can be simpler or less liable to require an overlooker's attention when once adjusted. 372 MECHANISM OF WE A VING PART Of the so-called positive let-off motions some merely give automatic Aveighting without attempting to pay off warp regularly. Others deliver warp positively, and then by hanging dead weights upon it after it leaves the beam a fixed tension is maintained ; most of these are a combina- tion of positive and negative parts working in conjunc- tion, but without direct connection with the taking -up motion. Others, again, may be defined as self-contained motions for letting off* warp and drawing away cloth positively. From the large number of inventions in the market it is ^ /y////////// / /yy ///// /// ///////y/y///y///y'/y/y///^^ Fig. 206. only possible to select one, or at most two, from each class ; still, if they fairly represent the different principles involved, a good general knowledge may be obtained, and the study of minor alterations can be safely left to the student. Charles Schilling of New York claims to be the inventor of a motion belonging to the first-named class. He uses ordinary weights and levers ^, a (Fig. 206), but fastens the weights to an endless rope c, that passes over guide pulleys fixed upon the framing, and also round rope wheel /. Compounded with / is a pinion g, the teeth of which gear with those of rack h suspended from the top of presser lever i. The fulcrum of i is at 7, and a weight at h Above MECHANISM FOR GOVERNING WARP 373 rack li a pressor acts against the under side of the warp, and as the latter is drawn away the presser rises, pulls up rack A, moves pinion r/, Avheel /, rope and weights h. It will be noticed that as rope ^; is crossed, any movement imparted to it will carry each weight h towards the ful- crum pin of its own lever, and thus the weight upon the warp beam is diminished in proportion as the diameter of nib [ 3 FiCx. 207. the beam decreases. This motion is not unduly complicated, nor is it difficult to adjust ; when a new warp is put in, it is only necessary to force the presser down and the weights will move to their proper places. If required, the weights can be removed during weaving, but they must be put back in the same places. Its defects are : that the warp is not delivered, but pulled off ; that the presser prevents the beam from readily taking back any excess of warp drawn away ; that no attempt is made to maintain an even tension during shedding; that the form of the presser is more liable to 374 MECHANISM OF WE A VING PART glaze the warp, and to cause the surface of the beam to be irregular than if a simple roller, extending from flange to flange, was used. Messrs. Hanson and Crabtree patented another automatic weighting motion, consisting of a beam a (Figs. 207 and 208), resting in leather-covered journals h. The upper journal is pressed against the iron hoop of beam a by means 210), which is fixed on one end of the warp beam and gears with worm h ; the latter is secured to the bottom of a short vertical shaft, and ratchet wheel c is secured to the top. This shaft also serves as a fulcrum for a bell-crank, or letting-ofl" lever, on arm d of which pawl e is hinged, and takes into the teeth of c. A spiral spring, hooked into c/, and also into a bracket on the framing, keeps the lever in one position when not otherwise controlled. An end of rod g is fastened to a slay sword by a stud, the of a flat spring c to give the necessary pressure, a portion of which is taken off every time the beam revolves by a peg screwed into the beam hoop, that comes into contact with star wheel and moves it one tooth, thus setting in motion wheels /, ^, A, worm i, worm wheel y, and causing screw h to ascend and slightly reduce the compressing force of c upon h. Fia/208. A motion belonging to the second class, as made by Smith Brothers of Heywood, consists of worm wheel a (Figs. 209 and XVIII MECHANISM FOR GOVERNING WARP 375 other is slotted, furnished with an adjusting screw, and by means of a second stud is held against the under side of arm / of the letting-ofF lever, so that, at each forward stroke of the slay, arm /is pulled far enough to allow pawl e to take a tooth in wheel c and drive the warp beam by G Fig. 209. worm h. But it is obvious that a tooth taken when a beam is full will deliver more warp than when the beam is nearly empty ; for this reason negative parts are added to regulate the supply ; these are : shaft A, vibrating in bearings and carrying a three-armed lever ^, 7, ^, near one extremity, and a two-armed lever near the other, i supports a warp roller I at its upper end, j a heavily weighted stalk m, and k carries a slotted rod ??, which is connected by a stud to the 376 MECHANISM OF WE A VI NG PART upper surface of arm / of the bell-crank lever. Rod n slides on / until the inner end of its slot comes into contact with the stud, where / is practically locked. The warp passes inside shaft A, over vibrator and forward to harness and fabric ; at the same time the weights upon stalks m keep it under a constant tension. Ratchet if moved one tooth for each revolution of the crank, would deliver considerably more warp than the taking- up roller requires ; when this occurs, the vibrator roller / F Fig. 210. is pulled back by the weights on and in falling back rod n is pushed forward by arm k until the stud in j touches the end of slot in n ; then, owing to the spiral spring being unable to draw back pawl e far enough to droj) over the next tooth in ratchet c, letting-off is suspended. The motion is fairly satisfactory for heav}^ work, but a greater strain is put upon yarn in an open than in a closed shed. In 1883 Keighley of Burnley contrived a self-contained let-ofl* motion, in which the beam is only weighted to prevent warp from becoming entangled by unwinding too XVIII MECHANISM FOR GOVERNING WARP 377 freely. After warp leaves the beam it passes over a heavy, cloth -covered presser roller a (Fig. 211), and under and nearly round a measuring roller />, having exactly the same circumference as the taking-u}) roller, viz. 15'', thence over vibrating bar and forward in the usual manner to the cloth roller. The centre of roller a is placed higher than those of roller h and pulley c, to allow the former to gravitate towards and thus, by their nip, to prevent the warp from slipping. All the parts of Keighley's motion are driven by worm which is set-screwed on the end of the bottom shaft, and is in contact with a wheel e secured to a short horizontal clutch shaft. Slightly in advance of e bracket / carries a flattened projection into contact with one of ten teeth on the side of clutch disc g ; the latter slides upon a key-way whenever the loom stops, and letting-ofl" and taking-up are disengaged in the following manner : — g is traversed later- ally upon its shaft by the oscillation of lever s on fulcrum and by a pin u entering the ring groove in clutch g. A rod V connects s with a three-armed lever having a horizon- tal arm fixed to touch spring handle y and a hinge at so that by turning it up to the perpendicular, the loom may be set in motion without delivering warp, or drawing cloth forward ; but if left horizontal when %j is shifted to the retaining notch, g slides into contact with /, and all parts of the machine act. If the weft breaks, cracks are prevented by the loom starting in advance of the clutch ; this is owing to the time taken by / in moving from tooth to tooth in g — say, on an average from 2 to 3 picks. A protruding piece 4 on 5 passes beneath a weighted lever z and invariably elevates it with .9, but z is free to rise alone ; z carries brake clog 1 at its inner end and is connected by rod 2 to a lever .3, shown resting upon weft fork lever 5, PART XVIII MECHANISM FOR GOVERNING WARP 379 hence, when y is moved outward, z rises with 2, and clog 1 leaves the periphery of fly wheel 6 ; the opposite movement of y brakes the loom. Measuring roller h is driven from the clutch shaft by wheel A, which should con- tain 15 teeth if the teeth in wheel i are to equal the number of picks per \ inch ; 30 teeth if they equal picks per \ inch ; 45 teeth for picks per \ inch ; and 60 teeth for picks per inch. Wheel i is thus seen to be the one to change when alterations are to be made in the number of picks ; it is attached to the rear end of a side shaft and is driven by carrier h. A worm / drives measur- ing roller and another wheel on the forward end of the same shaft drives through a carrier the tak- ing -up roller wheel, which contains 100 teeth; the latter is fixed upon a short shaft j?, revolving below and at right angles to the taking -up roller g, it is com- pounded with a worm which takes into the teeth of worm wheel r on q. It is well known that more than a yard of warp is required to weave one yard of cloth, but the exact amount varies with the weave, the thickness and closeness of warp and weft. Although it may be possible to ascertain by calculation the allowance for contraction, yet in practice it is a matter which is determined by experience. Assuming the percentage of warp-contraction has been obtained, the side shaft wheel must have one tooth less for every per cent of contraction than the taking-up wheel. The parts described above secure regular delivery of warp ; it now remains to be seen what provision is made for maintaining a constant and regular tension whether a shed is open or closed ; this consists in making the outlines of shedding and vibrator tappets exactly similar, but dilFering in throw. After making the following experiment, the 38o MECHANISM OF WE A VING PART writer is of opinion that the inventor has succeeded in his attempt : — A thread was passed between the pressing and measuring rollers, over the vil)rator, through healds and reed, and a light weight was attached to the forward end and permitted to hang loosely over the taking-up roller, the healds were moved up and down, and it was found that the weight remained stationary. The chief fault of this motion appears to consist in the number of parts added and in the trouble entailed in gaiting a new warp. One serious objection to positive motions is that an increase of moisture in the atmosphere causes warp threads to contract in length, and since they are held perfectly rigid by the cloth and warp beams, the tensile strain put upon them is often sufficient to pull fibre from fibre, if the loom is left stationary for any considerable time, and a large number of breakages occur before the warp thus exposed can be woven up. PAET XIX TAKING-UP MOTIONS In weaving by power, some means must be adopted to draw the fabric forward regularly as it is woven, else it would be impossible to produce goods having any approach to perfection. Such a motion does not present any exceptional mechanical difficulties ; on the contrary, parts were added to what was little more than an experimental loom which perfected the principle, at least half a century before the power -loom became a commercial success. The parts XIX TAKING- UP MOTIONS 381 employed were two rollers pressed tightly together and driven at a constant speed, the fabric passed partly round and between them, and was by this means drawn away at the required rate. But the foregoing is only a partial statement of the functions of parts required to govern the movement of cloth. It is of the utmost importance that uniformity shall be maintained ; and since after unweaving, and after weft has given out, the fabric must be let back, some efficient means of determining the extent of such movement is necessary. Or, in other words, some method of accurately fixing the position of the cloth fell is still a desideratum, for at the present time no satisfactory solution has been reached. To-day there are two classes of taking-up motions, known respectively as " positive " and " negative." The former may be subdivided into : (1) intermittents ; that take up as the slay falls back ; ^, that take up as the slay moves forward ; those where a driving wheel is changed ; and those where a driven wheel is changed. (2) continuous ; in this case the gearing is such that cloth is constantly drawn away if the loom is in motion. Of negatives, there are those that are self-contained, and those that act in conjunction with a loose reed. If a positive motion implies the use of parts where nothing is left to chance, then the one in most extensive use will by no means stand the test, for the cloth passes partly round a wooden roller, having, as a rule, its surface covered with a strip of thin metal which is first punched full of small holes to roughen one side, and is then wound spirally upon the roller, with the rough side exposed. Occasionally an iron roller is grooved longitudinally and transversely to give roughness ; but sand, emery, and pins driven into a leather foundation, are all used for the same 382 MECHANISM OF WE A VING PART purpose- -namely, to create friction, and so enable the roller to grip a fabric firmly and draw it forward. The bearings of beam a (Fig. 212) are in the end framing, and so placed that the periphery of a partly projects beyond the front edge of the breast beam ; it is also parallel with the latter, with the harness, and the back rest. Fig. 212. Beam wheel h is set-screwed on one end of the beam shaft ; its teeth gear with those of stud pinion c, and the latter is compounded with stud wheel both so named because they work loosely upon a short adjusting stud that passes through and is bolted to a slotted bracket, the slot being concentric with beam a. Wheel d gears with change pinion XIX TAKING- UP MOTIONS 383 and upon the same short shaft ratchet / is fixed. Two pawls or catches i rest freely upon the surface of/; g is simply a holding catch to prevent the ratchet wheel turning in both directions; it is keyed upon a rod Ji, extending across the loom, h carries at its extremity farthest from g a finger that presses against the weft fork lever ; the effect of this contrivance is to lift catch g above the teeth of / when weft breaks, and thus prevent taking-up. A driving or pushing catch i is pivoted on lever k ; the latter oscillates on centre and a slot in its lower arm allows stud m to be passed through. As m is bolted to the slay sword, and adjustable, it follows that a swinging motion will be given to ^, and catch i will drive ratchet / forward, / in turn will set the entire train of wheels in motion. The extent of motion in a ratchet wheel depends upon the position of stud m in the slotted lever. This is an intermittent motion, for it only draws cloth forward when the slay is moving back ; but by altering the form of catch its action will be to pull, instead of push, the ratchet wheel round, and cloth will then be taken up as the slay moves forward. In either case the pawl is the weak point of this motion, and irregularities are by no means uncommon, owing to its defective action. The catch often slips over a tooth ; it may take two teeth, or it may not be strong enough to take them regularly. A thin wooden roller 1 is pressed against the under side of a by two balance levers 2 acting at opposite ends of its gudgeons; they are supported on studs bolted to the framing and loaded at their lower ends ; hence a drives 1 by surface contact and winds on the cloth negatively. A taking-up motion requires to be so set that a pushing catch shall drop over a tooth of ratchet/ when the cranks are on the front centres, and to take one tooth at each movement ; 384 MECHANISM OF WE A VING PART if more than one tooth is taken, stud m must be lowered ; if less than one, it must be raised. Holding catch cj should clear a tooth by about Y when the cranks are on the back centres. A pulling catch will consume less power in drawing a fabric forward than a pushing catch, and it is less liable to slip, because taking -up is performed as the cranks move from the top towards the front centres; at which time a shed is either closed or only partially open ; whe reas a pushing catch acts as the cranks move from the front to the bottom centres ; or when a shed is quite open and the strain greatest. Set a pulling catch to drop over a tooth with the cranks on the back centres, and to take one tooth at each forward stroke of the slay, then fix the holding catch to clear a tooth about Y-> with the cranks at the front centres. Picks are placed closer together in one fabric than in another, hence the taking-up beam must be made to rotate correspondingly slower or faster ; this is done by making changes in the number of teeth in one wheel of the train, that wheel being generally the change pinion e. The selection of suitable wheels is a simple matter, especially if it is first clearly understood that, irrespective of their form or position, wheels are drivers, driven, or carriers. A driver is one that increases velocity in propor- tion to its increased diameter or teeth ; a driven is one that decreases velocity in inverse proportion to its increased diameter or teeth ; and a carrier merely conveys motion from one part of a machine to another without altering the value of a train ; such a wheel invariably works loosely and independently upon a stud or shaft. To find the number of teeth in any change pinion for a given number of picks per \ inch, multiply the teeth in all XIX TAKING- UP MOTIONS 3^5 driven wheels together, and divide by the teeth in driving wheel, multiplied by the number of \ inches in the circumference of taking -up roller, and by the required hx dxf picks per 4- inch, or r-r-== pinion required. If cxax picks change pinions are required to give 16 picks per \ inch with trains 1, 2, the former consisting of beam wheel of 75, stud pinion 15, stud wheel 120, and ratchet 50 teeth respectively, together with a roller 15" in circumference, train 2 has a beam wheel of 75, stud pinion 12, stud wheel 100, ratchet 50 teeth respectively, and beam 15'' in circumference. = 31*25 teeth required. 15 X 60 X 16 2 = 75 X 100 X 50 12 X 60 X 16 : 32*55 teeth required. Only whole numbers can be used, therefore 1 requires 31, and 2 either 32 or 33 teeth. In practice, however, it has been found that cloth contracts in length between the loom and the counter, and it is usual to add li- % to the calculated teeth in a pinion for this purpose ; thus 31-25 + 11 % = say 32 teeth, and 32 -55 + 11^ % = say 33 teeth in the wheels to be used. Where a large number of looms similar in make are placed in the same mill, calculations are shortened in various ways ; frequently by finding a constant number, which, divided by the picks required, will give the teeth in pinion. The constant for train 1 = 75x120x50 15 X 60 = 75 X 100 X 50 12 X 60 " 2 c + 11% = 528. 386 MECHANISM OF WE A VING TART Then to find a pinion for 16 picks 1 = 507 16 -32 and 2 = -— = 33 teetli. 16 Inverse proportion may also be used to find the number of teeth in a pinion; thus if a 33 pinion gives 16 picks, what wheel will give 20 picks % 20 : 16 : : 33 : 26, the number required. Simple as changing a driving wheel is, it involves unnecessary risk and trouble — risk of putting the wrong wheel on, trouble in making the calculation. Many minor variations from the above - described motion are to be met with ; for instance, it may be found that an alteration of one tooth in a pinion will make too great a change in the fabric. In such cases it is not uncommon to employ a train of 7 instead of 5 wheels. A driven wheel may be changed instead of a driver, and when this is done calculations are seldom required, as such wheels are selected for the train that the number of teeth in a change wheel corresponds with the number of picks per inch in the fabric ; thus, ratchet 24, first pinion 36 (the first stud wheel is the change wheel), second pinion 24, second stud wheel 89, third pinion 15, beam wheel 90 teeth respectively, and cloth roller 15*05'' diameter. teeth in change wheel equal picks per inch. If the wheel of 36 teeth is replaced by one of 27 teeth, a change wheel equals picks per \ inch; if by 18 teeth, a change wheel equals picks per \ inch ; and if by 9 teeth, a change wheel equals picks per \ inch. The number of teeth in each 24 x 89 x 90 89 36 X 24 X 15 X 15-05 ^ 90^^' but 89 + 1-1- % = 90-3.-. XIX TAKING- UP MOTIONS 387 being clearly stamped upon its surface, mistakes are not of frequent occurrence (see also Keighley's positive let-off and take-up motions, pp. 376-380). The main difference between intermittent and continuous motions is found in the manner of driving. In the latter a side shaft a (Figs. 213 and 214), inclined to the floor line, is fixed at right angles to picking shaft and gives motion to the fabric by means of bevels c, worm worm wheel /, change pinion ^, stud wheel A, stud pinion beam wheel Fig. 213. and beam m. The velocity of such motion depending in this, as in other applications, entirely upon the wheels employed. It is essentially positive in action, as slipping is next to impossible, and nothing is left to chance ; it also possesses other advantages ; for instance, if a loom is turned backwards the cloth is moved back with it ; but in intermittents the pawl continues to act in the same manner, irrespective of the direction of other parts of a loom; the consequence being that cloth is frequently drawn forward Avhen weft is absent, and cracks are pro- duced, unless the weaver disengages the taking-up wheels, 388 MECHANISM OF WEA VING PART and thus lets back the fabric before recommencing to weave. Negative or drag motions do not all act alike ; still in most of them the warp is held tight until the fabric, pushed forward by the reed, pulls warp from the beam ; in doing which, the fabric is slackened, and a weighted lever and pawl act upon a ratchet wheel to draw the fabric forward without employing change wheels ; but close at- tention must be bestowed by the weaver upon the rela- tive weighting of warp beam, and weighted lever, or the decreasing diameter of warp, and the increasing diameter of cloth roller, will cause the fabric to be drawn away irregularly. They are chiefly employed for fabrics that would be injured by coming in contact with a roughened roller, or where weft is so irregular in thickness that a fabric made with it would be full of uneven places if drawn away positively. Fig. 214. A XIX TAKING- UP MOTIONS 389 of their appointed places ; again, as thickness of weft alone controls the taking-up, fabrics of uniform bulk are 390 MECHANISM OF WE A VING PART produced, and letting- back is unnecessary when weft breaks. They are, however, more difficult to set for open Fio. 216. than for heavy fabrics, and the difficulty increases in pro- portion to lightness. As applied to fustian and velvet looms the parts consist of a cloth roller a (Figs. 215 and 216), upon which the fabric is wound, and to which a worm wheel h is fixed. A short XIX TAKING' UP MOTIONS 391 shaft c carries a worm that gears with ^, and two ratchet wheels 6, /, both having compound holding catches ; at e consisting of two, and at / of three, all slightly differing in length. Shaft c serves as a fulcrum to a slotted lever g ; upon the latter are a stud, to hold the pulling catch and a pendent stalk loosely jointed near its forward end ; slightly above the base of a collar is placed to rest upon a slotted arm projecting from a slay sword m ; the lower portion of k passes through the slot, and its collar permits the stalk to be loaded with adjustable weights 71, for the purpose of pulling ratchet e forward by catch ]i. When the loom is in motion the slay's vibrations are transferred to arm which lifts stalk lever (/, and catch li ; the latter, hooking on a tooth of ratchet ^, keeps the other parts named suspended, until the reed, acting upon the accumulating weft, slackens the fabric sufficiently to enable the w^eights upon k to overcome any remaining resistance and pull ratchet e forward, then holding catches on f prevent any backward movement. To increase the number of picks per inch weight must be taken from k^ and to reduce them it must be added. In either case the exact alteration is made by trial. In silk looms it is a common practice to control the taking-up motion by a loose reed that swivels from a pin in the cap end ; it is kept vertical by board h (Fig. 217) and spring c, the latter presses one arm of lever d constantly against h. Points 6, / on d are respectively fulcrum pin, and an adjusting piece with a T-shaped head. All the above-named parts are attached to and swing with the slay. Immediately in front of / a lever g is moved upon a pin at g', and connected at g to catch shown resting upon the teeth of ratchet 1, which, being compounded with 2, drive another pair of compounded wheels 3, 4, and 4 gears when the reed is pushed back, for at such times head / forces catch li to advance, but as / moves back a helical spring / contracts, pulls h forward, and sets the train in motion. XX BOX MOTIONS 393 To increase the number of picks spring c must be tightened, and to diminish them it must be slackened, by suitably moving set-screw c , PAET XX BOX MOTIONS Several attempts were made to adapt the power-loom to the production of checked and other fabrics requiring more than one shuttle before a successful motion was obtained. Dr. Cartwright used a flat tray divided into compart- ments, each large enough to hold a shuttle, and he con- trived, by pushing and pulling it automatically, to move the proper shuttle in line with the picker. Another inventor placed a series of shuttles inside the segment of a circle, which was pivoted above the shuttle race and capable of a swinging motion that carried any shuttle into position. A third inverted the above by fixing the shuttle boxes outside the segment of a circle, and placing its fulcrum pin below the race board ; but to-day all such methods have given place to two principles of governing shuttles known as " drop" and " revolving " boxes. The former was invented and applied to the hand-loom by Robert Kay of Bury in 1760, but Diggle, also of Bury, was the first to successfully apply it to the power-loom. The latter was invented by Luke Smith of Heywood in 1843. Both are now found in considerable variety of detail, but of the two, drop boxes are most varied. The ever-increasing number of different box motions 394 MECHANISM OF WE A VING part proves conclusively that none of them quite meet manufac- turers' requirements. Experience shows us that a good motion must be positive, and so connected with other parts of the loom as to ensure it acting as they act, and thus render it impossible for the boxes to get out of time or rotation with shedding and picking if a loom is turned back for unweaving, or any other purpose. Provision should be made for stopping the loom or motion, in case the picker is not clear of the boxes when a change is made ; and if a box is not lifted to its proper position, else a positive movement will result in smashes. Any shuttle should be capable of being brought level with the picker at any time ; and the motion imparted to boxes should be as easy as possible, or during the short time allowed for making changes vibration cannot be avoided, and this neces- sitates a reduced speed of loom. Again, it is undesirable that shuttles should be driven out of their boxes if a loom is turned backward, for it is both annoying and trouble- some to a weaver to replace them before re-starting the loom. If from these data we examine the motions in most extensive use, it will be found that each is wanting in one or another particular. Negative Drop Boxes — Diggle's Chain In 1845 Diggle patented a box motion that for effici- ency and simplicity had no rival ; it at once became a great favourite, and held its place for many years, but is now being slowly pushed out by more modern appliances ; nevertheless, it will probably long continue in use on heavy, slow-running looms, and also on those of medium weight and speed where not more than two shuttles are required. XX BOX MOTIONS 395 furnished with two or more boxes at one end, it remains essentially as Diggle left it. 396 MECHANISM OF WE A VING PART In Figs. 218 and 219 motion is obtained from a pinion wheel (X, keyed near the ex- tremity of the crank shaft, which drives a slide wheel J, containing four times as many teeth as a ; thus if a = 20 teeth, then & = 80 teeth. Upon the inner side of 5 a flange or slide c is cast concentric with the peri- phery, but broken at two points exactly opposite each other, and from the centre of each break a stud d pro- jects. The slide serves the purpose of holding an eight- sided star wheel e station- ary, and the studs give it an intermittent motion, by taking into one of the notches and causing it to make \ of a revolution every second pick. Star e is sup- ported on a stud in the end framing, and has fixed at its centre an eight-sided chain barrel /, round which an end- less chain g is passed ; the latter is composed of links of varying thickness, the thinnest being thick enough to bring the top shuttle in line with the picker, and the others are thickened to bring Fig. 219. XX BOX MOTIONS 397 any remaining shuttle of the series into the same position at the proper time. The links equal \ of a revolution of barrel / in length, and are fastened together by pins li^ each pushed through the holes in two links, and kept in position by split pins. All pins li project beyond the links, and fall into the notches of barrel /; by this means the chain is prevented from remaining stationary if motion is given to /. Immediately above the barrel centre a bowl i revolves on a stud in lever y, and rests upon a link of chain /. Lever j is centred at \ and has rod / jointed to it at m ; the lower end of I is attached to lever 7^, fulcrumed near the floor at and a spear rod f connects lever n and the shuttle boxes A thick link will elevate bowl % and a thin one will allow it to fall, thus a vibrating motion is imparted to the shuttle boxes. If for any purpose it is desirable to shift the boxes by hand, slide h must be taken out of gear with pinion a; this is instantly done by moving a lever r, mounted on a pin fixed to the frame ; in r a stud is secured to take into a ring groove in the boss of slide and a notched bracket s holds r in its ordinary position. Theoretically any shuttle can be instantly moved in line with the race board, but in practice not more than one box can be skipped in either direction, for the action being negative, the boxes fall by gravity, and the farther they fall the greater the rebound — a fatal defect in high speed looms. The chain is ill adapted for long patterns ; it is heavy, costly, and frequently becomes almost unmanageable. 398 MECHANISM OF WEA VING PART Smith's Modification In 1858 Mark Smith introduced a most important modification of Diggle's chain, which is an excellent arrange- ment for producing great variety and length of patterns with a short box chain, but does not touch other defects Fig. 220. in the original motion. Smith retained every part of Diggle's invention, but added pieces to permit of a link remaining stationary until a change became necessary. The parts added are a sliding block a (Figs. 220 and 221), fitted loosely on the slide wheel stud, but kept constantly rotating by pieces passing through holes in wheel n and entering grooves in its boss. The pieces so protruding XX BOX MOTIONS 399 carry ordinary pins / for driving the star wheel and a sliding motion allows them to be withdrawn out of reach of the notches. It thus becomes a question of moving the chain barrel o frequently or at long intervals. This is done by a vertical lever forked at its lower extremity to take into a ring groove in the boss of a. Arm I) is free to vibrate on a stud in the upper part of the framing, and has near its centre a bowl rolling against the face of a cam tZ, formed on one side of the slide wheel : c is thrust out by II --. I This Card vvorks the Boxes, as above, and at the same time acts upon the Lag Barrel. OOP o o This Card makes no change in the Boxes, but acts on the Lag Barrel. 00 00 This Card makes no change either in the Boxes or the Lag Barrel. OoOoO Fig. 237. 422 MECHANISM OF WE A VING For Six Boxes PART From I bo;i to 2 box. 0000 From 2 box to 3 box. 3 » » 2 4 » ..5 5 » ..4 00 OO^ From 1 box to 4 box. 4 » ..I 5 3 .» 6 „ „ 2 „ 6 „ oooo^ From I box to 3 box. 3 m m I » 4 » M 6 6 „ „ 4 „ I 00 O From 2 box to 6 box. 6 „ „ 2 „ 3 » » 5 5 m -.3 o 001- From I box to 5 box. 5 » « M 2 „ 4 n 4 » 2 000 From 1 box to 6 box. 6 „ „ i „ 3 » 4 4 ..3 .. o o Reverser to be fastened to the end of any card required for reversing. Fig. 238. XX BOX MOTIONS 423 A cylindrical bar parallel with the needles, rests upon a lag ; it is bolted to a bell-crank lever and as the lags elevate hj\ the vertical arm of x imparts a lateral movement to the shaft of pawl lever and so takes the upper pin out of gear with star and puts the lower pin in gear with a second star y ; the latter is connected to chain barrel 0 by pinions, one on the barrel shaft, the other on the boss of y. As a consequence, barrel 0 will turn in one direction if star t is the driver, and in the other direction if y drives. The motion is made for four and six shuttles, and the difficulties experienced in laying and reading a pattern chain for eccentrics, as compared with Shaw's, are clearly shown in the accompanying Figs. (237 and 238). Circular Boxes Although circular boxes have not met with so much favour in Lancashire as in Yorkshire, evidence is not wanting to prove that they are steadily gaining ground. The shuttles are arranged in chambers formed in a wooden cylinder, and are moved forward or backward as re- quired, to take each colour into striking range of the picker. Six shuttles are in general use, but twelve are not uncommon. In most cases a single box is moved, some are, however, capable of skipping one or more shuttles, but the system tends to complication and imperfect working. The cylinder a (Figs. 239 and 240) is supported by a hoop h at the forward end, and by a spindle c at the outer end. Upon c a disc d is fitted containing as many pegs, arranged in circular form, as there are boxes, and also a star wheel 6, against the under side of which an 424 MECHANISM OF WE A VING PART ordinary spring hammer presses to hold the boxes true. Two long hooks /, g are forced by flat springs h into contact with the pegs, and engage at opposite sides of disc d ; they are centred on bent levers % each with its Fig. 239. fulcrum pin near the bend. At the other end of i two vertical hooks j are hinged, and they pass through slots in levers h and I. Lever li is oscillated on a pin in the framing by a cam m on the bottom shaft, that turns against a bowl n. At 0 a knuckle joint is made, and a spiral spring XX BOX MOTIONS 425 stretched above it is strong enough to keep the lever straight, except when a shuttle is trapped ; in which case the joint gives Avay to prevent a smash. An L-shaped bar / rests upon cam 11 on the bottom shaft ; it is held against the framing by a bracket, and supports an octagonal card barrel s. The full part of u pushes up the barrel, but it is pulled down by a spiral spring v fastened to bar / above its bearing, and also to the fram- ing. A spring hammer keeps s steady, and a catch hang- ing from the upper framing, pulls at a star wheel x on the barrel shaft to give \ of a revolution to s as it moves down. Card barrel 8 is thus seen to have a vertical movement, and at its highest point a blank card pushes up two vertical feelers ([ on the bell- crank levers that are imme- diately above the pattern chain, and thrusts back hooks ,/ from the solid part of lever h ; the latter in its next upward movement misses 7, and the boxes are unaltered, but one hole in a card permits a feeler to enter a perforation in barrel 5 ; and this leaves g stationary, with hook i over the solid part of k, consequently both go up together ; they elevate the rear and depress the forward end of % and thus pull down hook / or g to turn the boxes. The parts, % j, /, q are all duplicated, so it depends entirely upon which feeler is left horizontal whether the boxes will move forward or backward. Fig. 240. 426 MECHANISM OF WE A VING PART A spring hammer is not sufficient to prevent the shuttle boxes from moving too far, hence a stop of some kind is invariably added. One maker uses two small catches each swivelling on a pin fixed below and at opposite sides of the boxes ; both are pulled apart at the top by a helical spring ^, that connects them at their lower extremities. Stops y rest against the inner curves of the drawing hooks /, until either of the latter is brought into operation, when it thrusts y into the path of the pins on disc and checks the boxes at the proper place. An attachment is found on most revolving box-looms for stopping the pattern chain if a weft breaks. One con- sists of an extra arm 1, fitted on the finger rod (see Inter- mittent Taking -up Motion), 1 passes through a slot in upright 2, and 2 supports a horizontal piece 3, beneath bell- crank levers I. Immediately weft breaks, the finger rod oscillates, 3 elevates I and pushes j ofi" lever \ thus stopping the boxes. A curved plate is fixed at each end of the boxes to keep the shuttles in position when not in use, and two cone-shaped rollers push the outer ends of the shuttle tips back as the boxes move. Without some such contrivance the boxes could not be moved freely at all times. Compared with drop boxes, the advantages are, that an increase in the number of boxes does not necessitate a decrease in the velocity of a loom, for the movement being rotary, the weight on one side of the cylinder balances that on the other side, and leaves the power required to move them fixed, or nearly so, in all cases ; also as there is no rebound, high speeds are practicable. They have, however, disadvantages, which to some extent counterbalance the advantages. In the first place, XXI TEMPLES 427 a loose reed is essential, because an ordinary stop-rod finger and swell act on the side of a shuttle, whereas in a revolving box, its top instead of its side is presented at the point where the finger would act. For which reason it is manifest that heavy cloth cannot be made. Again, unless the checking apparatus acts suddenly, and with certainty, there is a tendency to carry the boxes too far. PART XXI TEMPLES The tendency for cloth to contract in width is princi- pally due to the tensile strain upon warp, but that upon weft also affects it. When warp threads from top and bottom sheds change places, they bear upon straight weft and produce a series of corrugations along its entire sur- face ; and since a bent thread is longer than a straight one, extra material should be given off', but as this cannot be done, owing to warp closing upon weft all across the piece at the same instant, it follows that a considerable pull is exerted by bent weft to contract a fabric in the direc- tion of its width. Temples are employed to counteract this tendency during the operation of weaving, and so prevent side threads from being broken by the reed, and the reed from being injured by the warp. Self-acting temples date back to Dr. Cartwright's time, but were so imperfectly constructed that nearly half a century elapsed before anything satisfactory was ob- tained, and even then they were not generally used, for many cotton fabrics continued to be woven within the 428 MECHANISM OF WEA VING PART last twenty-five years with wooden temples that were moved by hand. At the present time, however, it is exceptional to find a cotton loom without a self-acting temple. They are now made in such variety of detail that it becomes necessary to select a few types from the many for description. The following only will be dealt with : — namely, trough and roller ; one, two, and three roller side temples, inclined and horizontal rings. The trough and roller is by no means the oldest self- acting temple ; it was preceded by the nipper and the horizontal ring ; still it has been used for many years to hold out light and medium fabrics. It consists of a semi- circular iron trough or tube a (Fig. 241), that extends across the fabric, and has portions of the ends hollowed out to form bearings for roller h ; bolt holes are also provided for securing two caps c above the roller. Eoller h is about \\' in diameter; it is fluted throughout its length, with the exception of about ^' at the centre, and then the flutes on one side of the plain piece have a left- handed thread cut amongst them to form sharp teeth, whilst a right-handed thread is cut amongst those at the other side. A roller when fixed in position should turn freely without touching the trough, and the front edge of the latter is moved as near the cloth fell as possible without touching the reed. The whole is supported by two long springs bolted to the front rail of the loom, so in case of a trap, the temple will roll back without doing damage. Cloth, in passing over the trough edge, is deflected round the under side of roller a, and the maximum amount of bite is obtained by raising the front of a until the fabric is bent at a sharp angle ; but at the best this temple is not adapted for heavy work ; it has not XXI TEMPLES 429 sufficient grip ; it does not distend a fabric ; the race board must be lower than is necessary for most temples ; it covers about 2'' of the newly woven piece, and thus Fig. 241. prevents to some extent the detection of floats and broken warp. Still it is good for light and medium work, woven in loose reed looms, as it assists in throwing back the reed when a shuttle gets trapped in a shed. 430 MECHANISM OF WE A VI NG PART Single -Roller Temples are specially adapted to hold out light fabrics ; they are also fitted on medium and moderately heavy looms, with long sweep cranks, say T or more, in which the slay moves too near the breast beam to leave room for two or three rollers ; they are fitted at each side of a piece in a cast-iron box upon a rod a (Fig. 242), provided with t^vo holes, for studs h to pass through ; and the latter are bolted on the breast beam. Behind each hole a spiral sj^ring is pushed over a stud for the purpose of holding Fig. 242. the temple forward, and a split pin, or thin collar c, secures rod a upon studs h. Where space is of less importance, a broad flat rod is often fastened to long springs similar to those used with the trough and roller. For very light goods temple rollers d are made of box- wood, with steel journals, and a thin steel hoop at one end prevents injury to the wood. Brass and steel rollers are also common, having in most cases finely -pointed steel pins driven into them spirally ; many rollers are slightly conical in form, the taper being from f to in diameter on a 4" roller. A temple box has a movable cap, which, when screwed down tight, leaves a space for the fabric to enter on a level with the roller centre. As the fabric is drawn forward, it is bent over the teeth and held by them XXI TEMPLES 431 with sufficient tenacity to ensure good weaving and still leave the piece free from temple marks. Many temples are capable of preventing a fabric from contracting unduly without being able to distend it. Con- sidering that cloth invariably contracts between the reed and the bite of a temple, it is often desirable to stretch it slightly during its contact with the temple. One of FiCx. 243. several similarly constructed single -roller temples has a fixed spindle d on which a loose boss a (Fig. 243) turns ; a is in diameter, and has fourteen longitudinal grooves cut in its periphery at equal distances apart. A saw J, w^th fine teeth and smooth protruding shanks, fits loosely into every groove of a. At each end of the spindle a hollow collar c, with an inclined edge, is secured, and all the straight projections on saws h enter the hollow parts of c, and are thus prevented from falling out. As boss a rotates, 432 MECHANISM OF WE A VING PART one end of every saw presses against the cam-shaped edge of the inside collar and slides outward, and the teeth carry out the cloth at both sides simultaneously. Two-EoLLER Temples Two -roller temples, from 3|'' to long, and from to in diameter, are largely used in the coloured section of the cotton weaving industry. Both rollers may be of uniform diameters ; they may be conical, with a slope of from ^' length ; they may be placed parallel ; or they may converge towards the centre of the fabric ; thus, outside inside to from centre to centre ; they may converge towards the outside, so that if they are apart at the inside, they will be from to apart at the outside. Rollers are sometimes similar in material and shape of tooth ; at other times one is wood, or brass, whilst the other is iron ; they have teeth inserted and formed in every conceivable direction and shape. The temple box is of cast iron and provided with holes on the under side, to allow size and dirt to fall through, but the cap is brass, cast iron, or steel. In every case it is made to carry the fabric into contact with the front roller approximately at its centre, and a longitudinal tapering piece runs down the middle to divide the cap into two semicircles and press the cloth low down upon both front and back rollers. Three -EoLLER Temples If three-roller temples have two rollers in the box, and the centre one in the cap, the top roller takes the place of the deflecting piece. But if two rollers are fixed in the TEMPLES 433 cap, and one in the box, it is equal to working with the former upside down. A temple box should allow the front roller to be brought within ^' of the box face, and should be set as near as possible to the reed without touching, with a slope coinciding with that of the fabric and warp, be- tween breast beam and harness eye when the shed is closed, and as low down as possible without touching the slay. The fixings must provide means of lateral adjustment, also of moving temples forward or back. Spiral springs, short horizontal flat springs, and long vertical flat ones, are in use for holding the temples forward and permitting them to move back when they strike the shuttle. Of the three, probably long vertical springs are best ; for they are not so liable to become stiff and spoil the reed as either of the others. Inclined Rings Inclined ring temples are almost as varied as horizontal side rollers. From a single ring, with two or three lines of pins, to upwards of 20 rings, each with one row of pins, are met with. A few are made with two rings fitted \Y apart on separate studs, placed one behind the other, and covered with a double semicircular cap. But by far the most general form of this temple is that of a series of parallel brass rings a (Fig. 244), with a single line of fine radiating steel pins. All rings are kept apart by a washer 6, flat on one side, but furnished with an eccentric boss on the other, equal in length to the width of a ring, and every washer has a hole drilled through it on the skew. A stud c is securely fixed in a thick inner end-piece ^, and rings and washers are dropped over it, care being taken to 2 F 434 MECHANISM OF WEAVING part keep the full side of each eccentric uppermost. When all are in position a second end-piece e is slipped upon and the whole bolted into the temple -holder /. Skew- drilled holes in washers have the effect of holding all rings a more or less diagonally to their axes, and the eccentrics carry all pins above the upper surfaces of h and inside their lower surfaces. Fig. 244. For rollers in diameter, rings a are often \' apart and slope approximately at an angle of 22° from the vertical. Broad looms, however, are provided with ring temples varying from ^" to 9" in length, and each contains 15 to 20 rings. As a rule, rings in long temples are not equidistant, nor is the angle a fixed one. A 17-ring temple, for example, may be \^~' in diameter, and have the two outside rings apart, whilst the two inside ones are \' XXI TEMPLES 435 apart ; also, as they approach the selvage, the inclination of each increases. These temples are, generally speaking, excellent for holding out and stretching a fabric ; they are applicable to wide and narrow looms ; those with a single ring act on the selvage only, and are mostly used for fabrics that would be injured by temple marks showing in the body. Horizontal Eing Horizontal ring temples are the oldest make now in use. A slotted bracket a (Fig. 245) is bolted to the inside of the breast beam, and temple-holder 1) is secured in the slot of a. A thumb-screw c traverses the bracket and goes through the piece in the slot, so by turning c to the right or left, temple plate d can be adjusted. Holder & is a thin slotted plate, with two overlapping edges, and has a flat spring riveted at its centre. Temple plate d is pushed between the overlapping edges of h and above its flat spring ; the combined pressure of spring and bent edges holds the former tight. A brass roller ^, If' in diameter, has three lines of radiating steel pins long ; it is fixed upon holder d^ in a horizontal position, by a screw passing freely through its centre, and is encircled by a rim /, in which a diagonal slit g is cut slightly beyond the centre of e and towards the outside, for the fabric edge to enter, and be bent over the pins li. A wider slit of ^' at the rear permits the piece to leave them. In action the selvage of a fabric is gripped and stretched between the point of contact and outer centre of roller ^, then, after passing that point, it contracts again until it finally passes out by the wide slot. 436 MECHANISM OF WE A VING PART In CclS6 cl shuttle is trapped, plate d is forcibly driven back ; but it requires to be set forward again by the weaver ])efore restarting the loom. This temple has good holding powers ; still it is too rigid, and somewhat troublesome to manipulate after unweaving. It is also dangerous to use on fast -running looms, on account of weavers c ) c Fig. 245. getting their fingers crushed when breaking off loose weft as the loom is in motion. PART XXII CENTEE AND SIDE SELVAGES Centre selvages are required to hold the edge threads in position when two or more narrow pieces are woven side by XXII CENTRE AND SIDE SELVAGES 437 side in a broad loom. Although such a selvage is inferior to a true one, it is, nevertheless, very serviceable, and is in extensive use on fabrics in which strength is of minor importance. It consists in twisting the outside threads of adjacent inner edges half or wholly round one or two ad- joining threads that may be stationary throughout, or interlace with weft in plain order. The simplest apparatus consists of a loop of worsted heald twine, with a smooth pendent mail at the bottom ; this is tied to the front shaft, and has the crossing thread drawn through it after bending the latter under one or more straight threads which are to be all drawn into one dent of the reed ; the crossed thread is also passed through an eye in a back shaft in the usual manner. Hence it will be lifted each pick — namely, straight by the back, and crossed by the doup or supplementary leash. From two to four dents must be left empty between each pair of selvages according to the fine- ness of the reed. The chief objections to the above plan are, that crossing threads can never rise much more than half a shed, that the leash chafes as it is pulled round the stationary threads, and breakages are frequent. Split fabrics are sometimes severed in the loom by draw- ing the gaps over sharp fixed knives, but more frequently after they leave it. Briggs Bury of Accrington, in 1889, introduced a pair of flexible chains having a small ring at each end ; the upper rings are attached to the back and front shafts respectively by heald twine, the chains hang free at their bottom ends, and all crossing threads are drawn through two rings, one on each shaft. When the front shaft is raised, a cross shed is formed and the chain on the back shaft hangs loosely ; also when the back shaft is raised the same edge thread is 438 MECHANISM OF WE A VI NG part xxii lifted in the open shed, but the chain on the front shaft is curled round the other side of the stationary thread. When both shafts are level both chains hang slack and form loops below the warp. Such chains by their superior strength are capable of resisting the saw -like action of the rubbing surfaces for a much longer period than twine, and their flexibility, cheapness, and weight, render them well adapted for the purpose. In 1886 Shorrock and Taylor patented a simple and efficient split motion, shown in front and side elevations in Figs. 246 and 247. It consists of a framing a, that is bolted to any convenient cross rail above the warp ; from a wire h descends to a point below the bottom shed line, and is there coiled to form two large eyes c, that take crossing threads 1 for adjacent selvages. At d two small flanged rollers turn freely upon studs in a. Straps e are bent round rf, led through guides /, and, on plain looms, one end of each is made fast by cords and a strap to the periphery of a pulley ^, set-screwed upon the heald roller shaft A. The opposite ends are secured to a piece of elastic i, which is hooked into the bend of a wire support y, rising from frame a. At / two smooth brass eyelets are fixed in e to take the crossing threads 1, and stationary threads 2 are drawn through eyes h in frame a. The oscillating motion of li causes cj to alternately un- wind and wind strap upon its surface. If the former takes place, elastic i contracts, draws up strap and thus carries eyelets /, with crossing threads 1, round flanged rollers and up on the other side of stationary threads 2, as shown in the sketch. The reverse rotation of li moves all the parts back to the starting-point, and places threads 1 on the opposite sides of threads 2 ; wires m merely serve as Fig, 246. Fig. 247. 440 MECHANISM OF WE A VING PART guides to hold ordinary warp threads out of touch with moving straps e. This split motion is comparatively inexpensive, and may be easily applied to tappets, dobbies, or Jacquards ; but in case either of the last named is used, jacks of one, and har- ness threads from the other, should be connected to a light lever, fulcrumed in such a manner as to multiply the lift of straps for they must in all cases move through consider- ably more space than healds or harness. A motion, apparently of French origin, is constructed on somewhat similar lines to Shorrock and Taylor's ; but there is this difference, that whilst the last named holds straight threads stationary, the one under consideration causes them to weave in plain order. To make the descrip- tion as perspicuous as possible, the mechanism will be dealt with separately as if it served two functions — one to weave gauze, the other plain. Figs. 248 and 249 are respectively front and side eleva- tions : a, &, c are three compounded pulleys suspended from a top rail of the loom, d' is a strap which goes partly round, and is screwed upon a; its ends are connected to two treadles that move alternately and cause pulleys a, h, c to vibrate, or one treadle and a spring may be substituted for two treadles. The thick line e is an endless piece formed partly of strap- ping fixed to the surface of c, partly of twine, containing two mails, /—one for each selvage ; it is led round warves or grooved pulleys g, h, g\ An oscillation of a, b, c in either direction will cause both mails /, and the warp threads they contain, to turn round warves g, g\ and ascend on the other side ; in doing this each mail carries its warp round two intervening threads that must be drawn into the same dent as their crossing thread. It yet remains to show how those threads are XXII CENTRE AND SIDE SELVAGES 441 moved in plain order. A third strap is secured to pulley Fig. 248. Firj. 249. 6, and its opposite ends have cords attached in the 442 MECHANISM OF WE A VING PART following manner : — Cord i is bent round warve ^, from whence it goes up diagonally to cord i', and is there tied. Cord % is taken round warve H and up again to i. Each is provided with two mails that form part of opposite selvages ; hence if an oscillation of l causes mails y, j to ascend, and /, / to descend, the motion of h when reversed will enable plain cloth to be woven by lifting /, / and sinking Taking both actions together, the cords appear somewhat involved ; but where it is simply a question of twisting a constantly moving thread partly round a stationary one, it is very evident that the parts are quite as simple as those of Shorrock and Taylor's motion. Inventors are continually adding to the number of appliances for w^eaving centre selvages, and it becomes a difficult matter to make a good selection from those available. Motions are often met with that twist one thread entirely round another in the form of a spiral. Probably two of the best known were introduced by Sir Titus Salt, and by Boyd. The first named has a split pinion a (Fig. 250) bolted on the crank shaft ; it is across the face ; it has sunk teeth, wide in the centre, and smooth flanges of equal width on either side. A wheel, or ring wide, has two smooth outer surfaces, and teeth projecting from the centre, for the purpose of gearing with the depressed teeth of a. The teeth on h are in the ratio of two to one in the pinion. King h has an inner diameter of 5|'', which is crossed by a spindle that is supported by pressing its ends into holes drilled in the ring. This spindle has a key fitted near one end, a pin hole drilled near the other, and a thread cut to a point somewhat beyond its centre ; it carries two lock-nuts c, having circular milled heads, two XXII CENTRE AND SIDE SELVAGES 443 spools 6?, a spiral spring and a short holding pin. Spools d are of brass, each over all by in diameter ; one has a key bed cut to receive the key in the spindle, the other would turn freely but for spiral which presses upon its inner end, and also abuts against a nut c ; hence by moving Fkj. 250. the nuts to compress the spring one spool will be thrust against the holding pin, and free rotation will thereby be retarded. Each spool has two threads wound upon it, and it is also necessary to fix them upon the spindle, so one will wind on, as the other unwinds yarn. One side of ring I is perfectly flat, but the other side has two curved flanges / exactly opposite each other ; both 444 MECHANISM OF WE A VING PART project about at their widest points, and within a space of 2 1'' they taper down to the ring. Below the centre of each flange a thread tube passes through ring h and pro- trudes on the flanged side ; both have an eye drilled in the sides that face the spools ; the threads from a spool enter one of the eyes, then one is taken to the right and the other to the left. Wheel h is situated above with its front edge immediately behind the healds, and its centre midway between the top and bottom lines of an open shed. It is held close upon a by a semicircular brass cap ^, bolted to any convenient bracket on a cross or back rail of the loom ; g is grooved to receive the teeth of &, and as a consequence the smooth edges of ft, &, g are all in touch. Continuous twisting is obtained as follows : — As h rotates, the flanges / are moved successively opposite the centre of a shed, and when this point is reached the two threads nearest the healds are distended by the flange and tube combined ; whilst the other pair, being farther away by the diameter of the ring, are closer together, therefore the separated threads pass on the outside, but after a half revolution of h the second pair are distended to assume the outer positions, and the first pair being contracted move between them. Care must be taken to insert doubling twist into the spool threads in such a manner that the rotation of h will not untwist and break them ; or, two pairs of untwisted strands may be wound upon each spool, when every revolution of the ring will put a twist into both pairs. If Salt's motion is in proper working order an exceed- ingly neat and strong edge is obtained. Boyd's contrivance consists of two circular spools which fit loosely inside brass holders. All are placed in recesses on XXII CENTRE AND SIDE SELVAGES 445 opposite sides of a central plate, and spools and holders are retained there by pressure from two thin flexible metal plates. The top of each holder is bevelled to slope outward, and the under side to slope inward. A back plate holds what would in ordinary splits be the stationary threads ; but here they move up and down continually in opposition to the spools ; thus, if for one pick these threads form the bottom shed, and spool yarn the top shed, on the following pick they will change places, in doing which they must slide between the spool-holders and the middle plate, but in moving down again to assume their original positions the upper bevels will carry them between spool -holder and outer plates, therefore they twist round and round the spool yarn, and make one half-twist between each pick of weft. A tappet and spring give the vibrating motion. This also forms a good strong selvage, but it may be affirmed of both Salt's and Boyd's inventions that their cost precludes general adoption. Plain Side Selvages Plain selvages are made on sateens and twills in various ways without increasing the number of shafts. By means of a contrivance known as a boat, a selvage sufficiently like plain cloth to answer the purpose is readily obtainable. If, for example, a 5 -shaft sateen, 1 up and 4 down, is required with such a selvage, two boats, situated at opposite sides of the piece, are supported by and vibrate upon round pins bolted to the inside framing of the loom, behind the healds, and below the warp line. A boat consists of a piece of wood 2" long, that is flat at the under side, but curved on the top to provide suffi- cient material in which to drill a hole large enough to 446 MECHANISM OF WE A VING PART take the fulcrum pin. Two pieces of leather, each about \Y long, and wide enough to reach across the wood, are nailed to the under side and turned round opposite ends, then a series of reed wires are riveted in the leather, twisted half round and bent over at the top, as seen in front and side elevations in Figs. 251 and 252, where a is the wood, 1) the leathers, c, c the wire, and d the hole Fig. 251. Fig. 252. for the fulcrum pin. Selvage threads intended for shafts 3, 4, 5 are passed through the loops of and those for shafts 1, 2 go through the loops of c, then every thread from c goes over a flattened down eye in shafts 1, 2, and every thread from c goes over one in shafts 3, 4, 5. As a consequence, a lifted shaft will elevate one end of a boat and depress the other end ; the former allows its threads to rise to a top shed, but the latter pulls all its threads down to the race board. XXII CENTRE AND SIDE SELVAGES 447 By suitably drawing warp through boat loops and over heald eyes, twills, and other satins than the one given as an example, can be woven with perfect plain edges in some cases, and with only slight defects in others. Another method of weaving plain cloth selvages is equally simple, and does not greatly differ in cost. It is illustrated in Fig. 253, where a is a light lever resting by its own weight upon the top of a cam &, and from which a cord c ascends to control a series of harness couphngs d, that are 448 MECHANISM OF WE A VI NG PART connected by their top loops to a second cord e ; the latter is bent round a grooved pulley /, and at its lower end supports a similar series of couplings r/, whose bottom loops are made fast upon a stout elastic cord A, or upon several such cords. Lever a must be heavy enough to distend cords (J in case the thin side of h is uppermost, and thus reverse the positions of mails in g ; but when cam h lifts lever a the elastics g exert force enough to move the mails back to the former position. Hence if d contain all odd threads, and g all even ones, plain cloth will be woven. A similar plan, but with a single mail, is adopted to manipulate a catch cord for fabrics in which two or more picks are to be driven through the same shed, and where, without some such contrivance, the weft would be drawn back again with the return of a shuttle. PART xxm TIMING AND FIXING OF PARTS In the early days of power-loom weaving it was dis- covered that an iron framing was best adapted to meet the requirements of a loom — namely, great strength and com- pactness, combined with cheapness. Strength — in order to resist the shocks and vibrations resulting from several distinct sets of intermittent mechanism, which, although working in harmony, and deriving motion from a common source, act and react upon each other to a considerable extent, and compactness — is necessary to economise floor space. The framing consists of two ends united by longi- tudinal, and supported by transverse rails. Its design is XXIII TIMING AND FIXING OF PARTS 449 of great importance, for it not only determines the posi- tion and relation of one part to another, but it materially affects the cost and amount of production, by rendering it an easy or a difficult matter to adjust, repair, and manipu- late the entire machine. The cost of running a loom depends largely upon the distribution of strength through- out the framing. If it is not strong enough to sustain the frequently recurring shocks at the points where impact takes place, vibration and breakage will be transferred to other parts. From the front to the bottom centres of the cranks' circles, beating up, shedding, and picking are performed, and the suddenness of their action consumes a large amount of power ; but after passing the latter point little force is needed to carry the cranks through the remaining f of their circles. The tappet shaft sustains the greatest share of strain, which in shedding is upward, whilst in picking it is diagon- ally forward. Strain on the crank shaft is backward and slightly downward ; therefore that resulting from the con- joint action of these shafts will fall somewhere between the two centres, and the framing should there have its greatest strength. Its height should be such that the weaver can readily reach any part requiring attention. From 33'' in narrow looms to 38'' in wide ones is the usual distance between floor and breast beam ; and in Lancashire an addition of to 3" gives the height of back rest. The width or stretch varies with the nature of the material to be woven and the mounting of the loom. An ordinary calico loom is from 36" to 38" wide, but narrow looms designed for heavy work are 40" to 44'', and broad looms from 52" to 56" wide. 2 G 450 MECHANISM OF WE A VING PART The warp line is determined by the relative positions of back rest, breast beam, heald eyes, and lease rods. Upon this, as upon many other important matters, experts hold different opinions, especially as to the exact positions the parts named should occupy. Some advocate placing the breast beam and back rest in one horizontal plane, and (with a lease in which two threads are over and two under each rod) the first rod and cloth fell to be equidistant from the centre heald shaft, and then sinking all heald eyes below a straight line drawn from back rest to breast beam, so that an upward movement given to some warp, and an equal downward movement given to the remainder, will form equal angles before and behind the healds in the bottom shed, and equal but smaller angles in the top one. See Fig. 254, in which a is the breast beam, h the back rest, dotted line c the horizontal plane containing both, d the closed line of warp, /, i, 7, respectively bottom and top lines of an open shed, g fell of cloth, and A front lease rod. It will be seen that angle ^, c equals angle /, c ; also that XXIII TIMING AND FIXING OF PARTS 451 angle y, c equals angle c ; but triangle /, c is greater than triangle j, i, c. The angles being the same, equal strain will be put upon the warp on both sides of the healds, but more on the bottom than the top shed. As a consequence, all threads occupying the former position will be tight, whilst those in the latter position will be slack. This is done with a view to prevent warp threads, that pass in pairs through the reed, from running together in the fabric, and is assisted by partially or wholly opening a shed intended for the following pick before the reed drives the last one into position, then, during the operation of beating up, loose warp threads are separated from tight ones, and the cloth appears to be composed of threads all equidistant. In Lancashire, as already mentioned, it is usual to elevate the back rest about 1^'' above the breast beam, and sink the heald eyes as before ; but cover is regulated in the cloth by moving the lease rods closer to the back shaft to tighten a top shed, and nearer the back rest to slacken one. This plan gives larger angles behind than before the healds, and puts more strain upon warp in one part of a shed than another. If fabrics are woven face down, the back rest is dropped and heald eyes are raised, to leave the bottom shed slack. Inequality of tension, however, should not be taken too far, or corded cloth and broken warp will be frequent. The bevel of a race board depends upon length of sword, positions of rocking shaft and connecting pins, throw of cranks and warp line. The warp line of an open shed determines the length of swords, and as the latter are often perpendicular when reed and fabric touch, the position of rocking shaft is in such cases also obtained. Some maintain that vertical swords are best on account 452 MECHANISM OF WE A VI NG PART of a blow being more effective when delivered at right angles than from any other point. If this contention is worthy of consideration, a blow from a reed to be at right angles to a fabric should be given when the swords incline towards the cranks, because a fabric usually slopes from breast beam to healds, and this of course implies moving the rocking shaft nearer the front rail. Cranks for narrow and medium looms move the reed 4|'' to 6'', and in wide looms from 6'' to 12''. When the requisite size of cranks and length of connecting arms have been obtained, proceed to fix the former's position in the framing by the rule given on p. 338. Couple arms and swords, place the cranks on their bottom centres, and make the bevel of the race board similar to the bottom warp line ; but reed and shuttle box backs are vertical when reed and fabric are in contact. After defining the position of the crank shaft, proceed to fix the bottom one so that picking and shedding tappets can be readily adjusted, but do not make the driving wheels larger or heavier than necessary, else as they revolve the energy stored up by them will be expended upon the frogs of a fast reed, and the brakes of a loose one, each time the machine is brought to a stand. In looms designed to weave stout fabrics, the driving wheels are intention- ally made large, and heavy fly wheels are fixed upon the main shaft, to accumulate force enough to give steady turning as the cranks approach their bottom centres. Still the rule holds good that there should be a minimum of weight in all moving parts. As the teeth of wheels wear rapidly at the picking point, it is advisable to move both ^ of a revolution, and then key them upon their respective shafts again before the teeth are totally destroyed. All parts of a loom are regulated by the movement of XXIII TIMING AND FIXING OF PARTS 453 the cranks, because they receive power direct from the main driving shaft and transmit it to all parts of the machine. When they are on the fore centres, beating up takes place, and as movement continues, other pieces of mechanism are brought into operation. Although picking follows shedding, it is a common practice to set the former parts first, as it is customary to run new looms some hours before any attempt is made to weave with them, in order to see that everything is working correctly. For good picking, all parts at both sides of a loom should be equal ; the tappets and cones (if this pick is taken as an example) should be in contact all the way round, and force should be properly directed. A long shuttle box is also of great advantage, as it materially assists in preventing a shuttle from flying out by serving as a guide. First of all, the picking arms and straps are to be prepared and fixed. Unless care is bestowed upon this matter, an arm will put a twisting strain upon the spindle as its picker approaches the forward end of its traverse. A long arm will give too much lever- age, but a short one will damage the strap, the spindle, and the picker. In most narrow looms, if an arm moves through an angle of 40°, approximately 30° will be passed through before it is parallel with the framing, and the remaining 10° will cause it to slant inwards. To find its proper length, put the cranks on their fore centres, draw the arm over the middle of spindle, and the centre of picking strap, where it leaves the arm, should be over the spindle centre. The length of strap can be obtained by turning the cranks on their back centres and moving the arm backward and forward, when a picker should slide freely from end to end of the spindle without the strap being unduly slack. If, how- ever, it is afterwards found necessary to lengthen or shorten the strap, this can be readily done, either by adjusting 454 MECHANISM OF WE A VING PART the strap itself, or by unscrewing the top nut of the up- right shaft and moving the upper ring and arm in or out, but it must be remembered the time of picking is thereby slightly altered. Slow-running looms should begin to pick when the cranks are on, or a little past the bottom centres, but it is usual to pick 10° to 15° before that point is reached in fast-running looms. To set a pick, hold a picker against a box end, turn the shaft until the picker begins to move, then see if the cranks occupy their proper positions ; if not, loosen the bolts that pass through the curved slots in the tappet, move both shell and nose in the required direc- tion, and screw all tight again ; after which, continue to turn the shaft until the tappet nose and cone centre are in contact, when the former should be from ^' to -f-^' away from the outer extremity of the cone. If more force is required, move the tappet nearer the upright shaft ; but it is desirable that as little power as possible shall be trans- mitted to a picker. In some cases shedding and picking tappets are correctly timed with each other, but not to the slay, then any altera- tion in the time of picking should be accompanied by a cor- responding alteration in the time of shedding. This is obtained by loosening the crank-shaft bearings, lifting one wheel out of gear with the other, and moving the tappet shaft in the required direction, after which the bearings are again secured. See that the shuttles are equal in weight and sufficiently heavy to overcome any drag from the weft. They should not, however, be too large, for large shuttles rccpiire large sheds, and the latter, by putting great strain upon warp, cause numerous stoppages to repair broken threads, and it must not be forgotten that more time is needed to repair warp than to replace weft. XXIII TIMING AND FIXING OF PARTS 455 If a check strap does not prevent a shuttle striking against a box end, the cop will be thrown off, or if bobbins are used, the holding pin will be broken, and shuttles will rebound. When cranks are on their bottom centres, a picker should be about X' from the box end. Shedding tappets are fixed to act at different times to suit the kind of fabric to be produced. If cover is unim- portant, and a tappet is used on which i of a pick is allowed for dwell, a shed will be almost closed when the reed and cloth are in contact, but it must be fully open by the time a shuttle enters the warp ; therefore, to set such a tappet, place the cranks in their picking position, and turn the tappet in its working direction until a treadle is pressed to the bottom, and there fix it firmly upon the shaft. Where cover is required, put the cranks on the top centres, turn the tappet round until both treadles are level, then screw it up tight. In case a difference is made in the diameters of tappet plates, care must be taken to connect the treadle to the back shaft that is actuated by the largest plate. A rule obtains amongst overlookers to lift the back shaft of a calico loom when a pick is delivered from the driving side. If springs are used to lift or pull down heald shafts, test each one before attach- ing it, to see that all are equal in strength, by suspending them in succession from a hook. Hang a weight of from 4 to 7 lbs. on every spring, and select those springs that stretch to the same extent. Back rest, breast, and warp beams should be parallel with each other ; the warp should be wound on with an even tension, and level all across ; it should be neither wider nor narrower than required, or the selvages will stand oblique and be liable to get broken. Eopes or chains must be wound on ruffles to permit of regular slipping, and the 456 MECHANISM OF WEA VING PART weights require constant attention to secure a fixed pull. Set the pushing catch of a taking-up motion to drop into the hollow of a tooth in the ratchet wheel when the cranks are upon their fore centres, then raise or lower the driving stud until the catch takes one tooth as the slay moves back; turn the cranks to their back centres and fix the holding catch about ^ past the bottom of a tooth. In fast-reed looms stop-rod blades must clear the frogs hy \' when a shuttle is in its box, and they should be long enough to stop the slay when the reed is more than the breadth of shuttle from the fell of cloth — say 2^''. If a steadying spring for a loose reed is just strong enough to prevent the latter from vibrating as a shuttle moves across, nothing more is required from it. But when the cranks are on their top centres, a space of 2'' should exist between fingers and frogs, or sufficient room will not be provided for the fingers to oscillate and slide along the upper incline in case of accident. The fingers must hold the reed firm when it touches the fabric. To do this suc- cessfully, their tips must pass the front edges of the frogs about A weft fork tappet should move its hammer lever as the cranks pass round the front centres, and so long as weft is intact, the fork must clear the hammer head by ^\ and also pass through the grid without touching at any part. Shuttle boxes ought to begin to move soon after a shuttle has entered a box, so that all motion will cease before the picker acts for the following pick ; hence boxes may begin to change as cranks reach the top centres, for then half a revolution is allowed to move and steady them before the following shuttle is driven across. XXIV WEAVING ROOMS OR SHEDS 457 Temples are fixed nearer to or farther from the reed, as the piece to be woven is light or heavy. The inclination also varies, but as a general rule they are set as near the reed as possible without touching, with the front edge from to -^-J' lower than the breast beam. PAKT XXIV WEAVING EOOMS OR SHEDS Weaving sheds, in this country, are now generally built to provide for a large number of looms on the ground-floor, in order to secure freedom from vibration, uniformity of humidity and temperature in the atmosphere ; also, because the plan admits of the most convenient arrangement of machinery for the worker — easy supervision, small cost of building, gearing, and carriage of material from place to place. A situation protected from dry winds and open to moist ones, with a clay subsoil, is desirable. But other things must be considered when a site has to be selected, such, for instance, as the close proximity to a good supply of water and fuel, ready access to the market for manu- factured articles, and a locality in which experienced opera- tives are plentiful. In many cases preparatory machinery is placed in a contiguous building of two or more stories for the purpose of economising floor space, but whatever the plan adopted, machinery should be arranged to prevent the material going over the same ground twice, or the cost of working will be increased. Shed walls a (Figs. 255 and 256) are U'-O" to 16'-0" high and usually 14'' thick, except when one wall h has 458 MECHANISM OF WE A VING PART to support the main shaft c, then it is increased to 18" or 24'' ; but this shaft sometimes rests upon cohimns built in the wall, and it is desirable for the sake of cleanliness and comfort to the work-people to isolate the shaft from the shed by walling it off. The roof is supported partly by the walls, but principally Fig. 255. by a series of cast-iron pillars d erected upon a concrete and stone foundation. In sheds constructed to weave plain cloth, the pillars have a diameter of b" at the bottom, and \y' at the top, with a metal thickness of |" at the former and ){ the latter places ; as they only support a light roof and the shafting, this strength is ample ; but in places where Jacquards are used the columns should be stronger, XXIV WEAVING ROOMS OR SHEDS 459 and haVelipi (?9sl; uporv ifiem from ^^'6'' to 12''0'' above the floor line to support a network of light girders on which the Jacquards are to rest. An additional Y in diameter and y in metal thickness will in most cases suffice. The arrangement of columns depends largely upon the class of loom to be used ; they are generally arranged to take four looms — two in length, and two in width, and leave walking space between and round them. Narrow looms are from 3'*6'' to Y'iy wide, averaging i'-O" or 4c''V and broad looms are from 4''7|'' to 5'*8" wide, but may be taken at The working alley e should be 20'' to 22'', the back one / 16" to 18", and the end passage g 24" wide ; therefore, taking the narrowest looms, we have 3''6" X 2 = 7'-0"+20"+16" = 10' '0" from centre to centre of pillars in lines running from east to west. For a 4'-0" loom at least 11' '4" should be allowed, but in some new sheds filled with narrow looms as much as 12' "9" is provided. Broad looms with two or three warp beams require 14' "4" to 14' "6" space from pillar to pillar. Looms placed in the same shed are sometimes of such varied lengths that it is out of the question to attempt to fix them so that they will relatively occupy a fixed posi- tion to the pillars. Pillars in lines from north to south are, as a rule, 15'-0" to 20'-0" apart, but a new system is being- introduced by which their number is greatly reduced ; thus, instead of arranging them — say, ll'-O" by 16' -9" apart as in the old plan, they become 22''0" by 33'-6" apart. This necessitates supporting alternate lines of shafting from the roof, and strengthening both roof beams and pillars to guard against vibration. But it is obviously a much easier matter to arrange looms of different lengths in a shed of the new than in one of the old type, and it also facilitates the carrying of warps. Columns have a uniform arrangement, except where 460 MECHANISM OF WE A VING PART main passages run t^ntxiely through -the hui|oii«g ; these latter are sometimes against one, sometimes against both end walls ; at other times a wide passage runs down the middle of a shed. At least should be allowed for each. A top light is always used in weaving sheds, and the roof windows li run from east to west to give a northern light, which is the greatest, steadiest, best adapted for im- parting a uniformity of temperature, and preventing the sun from being troublesome. Eoofs are made of wood or iron, and each column is cast in the form of a shoe i to take horizontal beams j that run from north to south. These beams bind columns and walls together, and also support the roof. Cast-iron gutters k rest immediately over the pillars, but are at right angles to beams y, and are connected to parts Z, that make an angle approximating to 65"" with the water level; such parts carry the upper portions of the roof, the gable of which forms a right angle. The steep portion is glazed and serves to keep the snow from collecting upon it in winter, and prevents the sun from shining in too freely in summer. In large sheds it would be next to impossible to remove water fast enough during storms, if the only outlets were at the walls, therefore the hollow columns are utilised for this purpose, and drain -pipes are run beneath the floor to carry the water away. It is ques- tionable whether this system is the best, for in addition to the difficulty experienced in getting at the drains to clean them out, there is always a liability for water to find its way to shafting and bearings. Another plan for removing rain water from shed roofs is by fixing pipes under the gutters and parallel with the beams; they are small in diameter where water is first received, but in- crease in area at each succeeding bay, in proportion as the XXIV WEAVING ROOMS OR SHEDS ^ 461 volume of water increases, and small manholes are provided at intervals which allow the pipes to be cleaned out readily. The main driving shaft c is always fixed against that wall termed the gearing wall, and it should be placed 12'*0'' to 13'*0'' above the floor line to work by bevel wheels all line shafts m — namely, those that run parallel with and between alternate rows of looms n. In most cases the last-named shafting is at right angles with the window lines, in order that the slays shall not throw shadows on the warps, and thus obstruct the weaver's light. Still sheds are geared with the shafting and gutters parallel. A shed is usually adapted for a given number of double rows of looms, but this plan requires an extra shaft, and some manufacturers prefer to have one or two single lines of looms set apart for their least skilled weavers. Main shafts are driven by wheel gearing, by ropes, and less frequently by belts. Other things permitting, it is advisable to place the engine at or near the centre of the gearing wall, and elevate the engine bed until it is high enough to permit the power to be passed direct to the shaft by spur wheels 0, j9, or if the velocity is unsuitable, a short horizontal shaft is employed. The crank shaft of the engine and the main shaft are then parallel, and equal power will be transmitted from the latter on both sides of wheel p ; or ropes may be employed, instead of tooth gearing, by leading them from the rope wheel of an engine round a rope wheel on shaft c. Every length of shafting diminishes in diameter in propor- tion to its distance from the point where it receives motion ; its dimensions will depend upon the nature of the looms to be driven and the distance from support to support. Generally speaking, this shafting is stouter than that neces- sary for ordinary driving because of the very unsteady action of looms. Power is distributed to the various line 462 MECHANISM OF WE A VING PART shafts through bevel wheels g', the diameters of which should be such as not only will impart the required velocity, but will enable the teeth of one to work by degrees in all the teeth of the other ; a difference of one tooth is often suffi- cient to ensure this. Line shafts may be driven direct from the engine if the latter is placed with its crank shaft parallel to those of the former, and if half the ropes are taken to pulleys on line shafts at one side, and half to those on the other side of the rope wheel, then by means of similar ropes and similar wheels every shaft in the shed can be driven. A shed floor r is covered with heavy slabs of stone to give looms a firm base. Looms are made right and left handed, to bring the driving pulleys together, and are placed back to back; they are formed into groups of four, and two rows are driven from one line shaft which is fixed to the pillars, and further supported by hangers from one of the roof beams. Each group is either driven by one broad drum s (20'' to 24'' wide by 15" to 20" in diameter) capable of carrying four driving straps, or what is often better, by two narrower drums fixed as close together as possible on opposite sides of the shaft bearings, each driving a pair of looms, by pulleys from 8" to upwards of 18" in diameter, according to revolutions of line shaft and required speed of loom. Loom ends should be fixed parallel to give straight lines of machinery. In order that this result may be attained, the crank shaft of one loom is made longer than that of its fellow by at least the width of both driving pulleys. It is essential that all shall be parallel with the line shafts, hence measurements are made from them. The usual method of procedure is to drop a plummet from different points along the first shaft, and where it touches the floor to make a permanent mark XXIV WE A VING ROOMS OR SHEDS 463 with a chisel ; then to chalk a cord, stretch it over two marks, lift it at the centre a few inches above the flags, and let it go suddenly, w^hen a straight white line will be visible on the floor ; it must be made to extend the whole length of the building. Other similar lines may be made below alternate shafts, and afterwards a line is drawn at right angles by means of a piece of light wood 10' '0'' to 12''0'' long, through which a nail is driven near each end. The rod is placed longitudinally over a chalk line and the position of each nail marked. From one mark an arc of a circle is described on both sides of the line, then from the other two more arcs are produced to cut the first pair, and the points of intersection are used as centres through which a chalked cord is to be stretched to make the line sought. From these lines the positions of the loom feet can be obtained, care being taken to fix them so that the pillars will not obstruct the back alleys more than necessary. A long straight-edge is laid upon the loom in various posi- tions, with a spirit-level on the top of it, to see whether or no packing is required beneath the feet ; if so, thin pieces of flat wood are cut to the required size, the positions of the foot holes marked, the loom removed, and holes are drilled in the flags, into which dry wooden pegs are driven tight ; the loom is then put back into position, and long nails are driven through the feet into the wooden plugs, the latter swell with the moisture of the floor and securely hold the loom. Calculations for speeds are of the simplest description, therefore two examples will be sufficient to clearly show the method of procedure. First, to find the diameter of drum required to drive a loom at the rate of 200 revolu- tions per minute, if its pulley is ^" in diameter, and the line shaft makes 130 revolutions per minute — 464 MECHANISM OF WEA VING part xxiv 130 : 200 : : 9 : 13'9. Answer 13-9^ Second, if a line shaft makes 140 revolutions per minute, and is furnished with a drum 15'' in diameter, what diameter of loom pulley will give 220 picks per minute % — 220 : 140 : : 15 : 9 55. Answer 9-55". Many other small matters require close attention, but they can only be properly dealt with as they arise in practice. There are few industries, if any, of which it can be said with greater truth than of weaving, that success depends upon strict attention to detail. INDEX Action of Bonelli's Jacquard, 175 of a Jacquard, 133 of a loose reed motion, 322 of single lift negative dobby, 83 of stocks and bowls, 74 of stocks and bowls, with closed shedding motion, 74 of stocks and bowls, with open shedding motion, 79 of tappets, 25 of weft fork, 354 Advantages of a cross-border Jac- quard, 156 of Crossley's Jacquard, 144 of double lift dobbies, 96 of open shedding, 21 of oscillating tappet, 56 of Shaw's drop box motion, 411 of shedding with liealds, 3 of a twilling Jacquard, 168 of tying below a comber board, 189 of Verdol's Jacquard, 164 Aggregate motion in a shuttle, 271 Aim of inventors of modern picking motions, 265 Ainsley's Jacquard, 138 Altering speed of loom, effect on shuttle, 277 Antifriction bowls, 38 Application of cranks to the loom ,336 Archer's dobby, 86 Arrangement of columns in weaving sheds, 458 Attachment for stopping a revolving box motion when a weft breaks, 426 Automatic reading - in machine, attempts to produce, 241 2 Automatic repeater card - cutting machine, Devoge and Co.'s, 236 repeater card - cutting machine, M'Murdo's, 238 repeater card - cutting machine, Nuttall's, 239 shuttle guards, 329 warp weighting motion, Schil- ling's, 372 warp weighting motion, Hanson and Crabtree's, 374 Bannister harness, 203 Barrel for weaving cross borders, 47 tappet, 45 Beams and rollers to be parallel, 369 Beating up, 332 up by compressed air, 333 up by flat springs, 332 up, position of parts, 338 up, power of slay, 345 Bessbrook Jacquard, 167 Bevel of race board, 347 Blackburn dobby, 97 Bonelli's Jacquard, 172 Box motions, 393 motion, flat tray. Dr. Cart- wright's, 393 motion, inside segment of circle, 393 motion, outside segment of circle, 393 Boyd's centre selvage motion, 444 Brake, Haythornethwaite's, 359 ordinary, 357 Brakes, 357 H 466 MECHANISM OF WE A VING Buffers, 312 Building rods, 185 Bullough's trough and roller temples, 428 Burnley dobby, 107 Bury's flexible chains for centre selvages, 437 Butterworth and Dickenson's dobby, 107 Calculations for consumption of whip threads, 263 for order of knitting healds, 15 for power consumed in picking, 299 for rate of knitting healds, 15 for speeds of looms and shafting, 463 for speed of tappets, 59 to determine amount of eccen- tricity in a slay, 340 to find the size of taking - up wheel, 384 Cams to give an eccentric motion, 36 Card cradles, 250 cutting, 227 cutting, with punches arranged by hand, 228 lacing, hand frame, 247 lacing machines, Reid, Fisher, and Parkinson's, 248 lacing machines, the Singer Co. 's, 249 lacing twine, 248 saving Jacquards, 156 Cards to fall over loom end, 183, 187 to fall over warp, 183, 187 wiring, 249 Carpet pick, 288 Catch threads with mail, cam, and lever, 448 Centre of gravity in a shuttle, 267 shed dobby, 95 shed Jacquards, 138 shedding, 19 shedding compared with stationary bottom shedding, 20 selvages, 436 selvage motion of French origin, 440 selvage motion, Shorrock and Taylor's, 438 Centre selvage motion for continuous twisting, Boyd's, 444 selvage motion for continuous twisting, Sir Titus Salt's, 442 weft fork, 361 Centred tie harness, 195 Chains versus ropes for weighting warps, 369 Changing a driven wheel in taking- up motions, 386 Check-strap, 311 method of attachment, 312 Cheetham and Sutclitfe's open shed Jacquard, 152 Christy's positive dobby, 110 Circle frame, movement of, 317 Circles, 317 Circular boxes, 423 Clasped healds, 5 Classification of Jacquards, 121 Closed shedding, 18 Collier, Evans, and Riley's swell, 318 Columns in weaving sheds, 458 Comber board composed of slips, 180 board, solid, 180, 201 Combination of positive and nega- tive parts for delivering warp. Smith Brothers, 374 Comparative consumption of power when picking from the crank and bottom shafts, 288 Comparison of reeds, 351 of semi-open shedding with other methods, 22 Compound harnesses to save cards, 199 Jacquards, 162 reversing motions, 68 tie harnesses, 197 Cone pick, 296 Connecting harness threads to tail cords of Jacquard, 185, 189 Construction of cone-picking tappet, 302 of curved plate for lever pick, 282 of Jacquard lifts, 176 of lappet wheels, 254 of lappet wheels, Scotch make, 261 of negative tappets, 40 of positive tappets, 47 I of reed, 350 INDEX 467 Continuous taking-up motions, 387 Cords instead of Jacquard hooks, 135 Coupling, 179 Cover on cloth, 38, 450 Cowburn and Peck's box motion, 416 Cranks for beating up, 336 Cross-border Jacqiiards, 157 Crossley's Jacquard, 141 Davenport and Crossley's cross- border Jacquard, 157 Deadweight as a reversing motion, 62 Decked dobbies, 88 Defects of clasped healds, 6 of closed shedding, 19 of Crossley's Jacquard, 145 of dead weights for moving healds, 63 of eccentric drop box motions, 415 of Jacquard shedding, 192 of negative picking, 266 of ordinary brake, 358 of single lift negative dobby, 84 of springs as reversing motions, 63 of tappet shedding, 27 of weight and rope system of governing warp, 367 Depth of shed, to lind, 30 Design of end framing of loom, 448 Detectors for double cylinder Jac- quards, 148 Devoge's automatic card -repeating machine, 236 Jacquard, 141 stop motion for double cylinder Jacquard, 151 Diggle's box motion, 394 Dimensions of columns in weaving sheds, 458 of cranks, 452 Division of parts of loom, 3 Dobbies, right and left hand, 94 with horizontal movement in cylinder, 81 with movable needle plate, 88 with rotary movement in cylinder, 90 with swinging movement in cylinder, 89 with vertical movement in cylinder, 88 Dobby, Blackburn, 97 Butterworth and Dickenson's, 107 Christy's, 110 Hattersley's, 102 lags, 93 Lupton and Place's, 107 pegs, 93 positive open shed, Knowles's, 113 reversing motion, 109 shedding, 79 Ward's, 102 Double equal plain tie harness, 201 lift dobbies, 96 lift double cylinder Jacquards, 146 lift single cylinder Jacquards, 138 picking bands, 310 shed Jacquards, 160 Doup harness, bottom, 216 harness, top, 222 harness in two sections, 223 Drafts, 12 Driving shafts, and gearing of weav- ing sheds, 461 wheels of looms, 452 Drop box motion, Cowburn and Peck's, 416 box motion, Diggle's chain, 394 box motion, Honegger's, 400 boxes, Kay's, 393 boxes, Knowles's, 403 boxes, to economise pattern chain, Shaw's, 410 boxes, Shaw, 405 boxes. Smith's modification of Diggle's chain, 398 and revolving boxes compared, 426 Dwell of tappet, 37 EcCENTRrc drop box motion. Hacking and Co.'s, 411 drop box motion, Whitesmith's,411 Eccentricity in a slay to be of value in weaving, 338 Effect of coiling ropes round beam ruffles, 370 of connecting arms on movement of slay, 340 of dipping the slay and bending the reed, 304 of high speeds on a pick, 298 468 MECHANISM OF WE A VING Effect of tappet treadle on move- ment of healds, 42 of treadle bowl on movement of healds, 38 of weight on the movement of a shuttle, 267 of worn parts of fast reed motion, 322 on shuttle of weft coiled on the tongue, 270 Electric Jacquard, 172 Elliptical wheels to give an eccen- tric motion, Dr. Cartwright's, 333 Equal compound harness, 197 Essentials of a good pick, 274 Estimate of wasted force in picking, 274 Eyed healds, 7, 11 Farrell's automatic shuttle guard, 329 Fast and loose reeds, 318 Figuring harness, 179 Fixing comber board in j)osition, 182 of tappets on loom, 25 Flexible chains as doups for centre selvages, Briggs Bury's, 437 Force of picking, 288 Fork and grid, 353 Framing of loom, 448 Frogs, 320 Gauze drafts and ties, 209, 211, 219, 222, 225 harness for bottom doup, 216 Jacquard, 218 produced with beads, 211 produced with needles, 212 reed, 213 weaving, 209 woven with double lift shedding motions, 220 Goos's Jacquard cylinder motion, 127 Jacquard hook, 136 Greek healds, 4 Green and Barker's stop motion for double cylinder Jacquards, 149 Guides to a shuttle, 268 Hacking and Co.'s eccentric drop box motion, 411 Hahlo, Liebreich, and Hanson's spring- easing motion, 66 Hall and Son's semi-automatic guard, 327 Hamblet and Clifton's semi-auto- matic guard, 329 Hanson and Crabtree's automatic warp weighting motion, 374 Harker and Grayson's swell, 325 Harness building, 184 dressing, 191 tied above the comber board, 184 tied below the comber board, 189 Hattersley's dobby, 102 Haythornethwaite's brake, 359 Heald calculations, 12 making, 8 setting, 17 shedding, 3 sizing and varnishing, 8 twine, 7, 9 Healds, 3 order of knitting, 15 position of, 7 rate of knitting, 15 with mail eyes, 11 varying velocities of, 36 Height of Jacquard, 182 of loom framing, 449 Honegger's box motion, 400 Hooks pushed on griffe by blank in card, 86 Horizontal ring temples, 435 Howarth and Pearson's double-shed Jacquard, 160 Inclined ring temples, 433 Introduction of dobby, 80 Jacquard, Ainsley's, 138 Bessbrook, 167 Bonelli's, 172 Crossley's, 141 Devoge and Co.'s, 141 Goos's cylinder motion, 127 Lambert's, 139 Shields's, 167 Verdol's, 164 bottom l)oard, 131 INDEX 469 Jacquard cards, 121 cards, two patterns on one set, 162 catches, 125 cross - border, Davenport and Crossley's, 157 cylinder, 122 cylinder motions, 124 double shed, Howarth and Pearson's, 161 griffe, 132 heel rack, 129 hook, Goos's, 136 hooks, 131 lifting tackle, 141 lifts, 176 machine, parts invented by Jacqnard, 120 mails, 180 neck cord, 131 needle board, 129 needles for single lift machine, 127 open-shed, Cheetham and Sut- cliffe's, 153 shedding, 119 single acting, 122 spring box, 129 ties, 193 Jamieson's tappet, 43 KEiGHLEY'sletting-off and taking-np motions, 376 Kenyon's under motion, 67 Knowles's positive open -shed dobby, 113 Lappet feelers, 252, 258 needle frame, 256 pin frame, 257 shedding, 250 slay, 251 slay, Scotch make, 260 tension cords, 259 wheel, 253, 260 Leading hook of Jacquard, 187 Length of picking arm, 300 of picking band, 309 Letting-off motions, 365 Levelling apparatus for dobby, 117 apparatus for healds, 21 Lever pick, 280 pick for pick and pick motion, 283 Lift of tappet, to find, 30 Lifts for compound Jacquards, 179 Lingoe, 179 Link -saving motion for drop boxes, Cowburn and Peck's, 420 Locking motions for double cylinder Jacquards, 148 Long picking bands, 310 Loose reeds, 322 Lupton and Place's dobby, 107 Lyall's positive picking motion, 305 Mail, 180 Manual repeating card - cutting machines, 229 Marriott's semi-automatic shuttle guard, 328 . M'Murdo's automatic repeating card-cutting machine, 238 Metallic clips for heald eyes, 10 Method of determining the amount of eccentricity in a slay, 341 of determining the variation in lift of harness, 192 Methods of attaching picking bands to loom, 310 of checking rotation in shuttles, 268 of supporting swivel shuttles in frame, 314, 316 Mixed tie harness, 197 Modern system of arranging columns in Aveaving sheds, 459 Modification of Jacquard hook, 135, 136 of Jacquard needle, 135, 141 Modifications of Jacquard machine. 134 of the Bessbrook Jacquard, 169 of weight and rope warp govern- ing motions, 368 Motion of shuttle desirable, 276 Mounting thread, 182 Movement of a shuttle, 271 Movements of loom, 3 Nature of movement in healds, 36 Neck cords, 131 Needle board, 129 frame for lappets. 256 Negative dobbies, 81 or drag taking-up motions, 388 470 MECHANISM OF WEAVING Negative drop boxes, 394 drop boxes, Diggle's chain, 394 drop boxes, Honegger's, 400 drop boxes, Kiiowles's chain, 403 drop boxes. Smith's modification of Diggle's chain, 398 picking, 278 tappets, 40 Nuttall's automatic repeating card- cutting machine, 239 chain tappet, 57 Open-shed dobbies, 105, 113 shed harness, Wilkinson's, 226 shed Jacquards, 152 shedding, 21 Ordinary Jacquard harness, with doup added, 215 Oscillating tappet, 54 Over and under motions, 61 Over-picks, 296 Parts of a loose reed motion, 322 of a loom regulated by cranks, 335 used to move healds, 24 Piano card-cutting machine, 241 Pick, carpet, 288 cone, 296 Jackson's, 285 Lyall's, 305 revolving scroll, 293 stationary, 291 swivel, 290 Yates's, 290 and pick motions, 283, 287 to impart a push instead of a blow to the shuttle, 278 Pickers, 307 materials used in their manu- facture, 308 treatment after delivery at mill, 308 Picking, 265 arm, 265, 280, 286, 289, 296, 300, 453 bands, 309 from the crank shaft, 288 motion, Kay's, 265 Pin frame for lappets, 257 Placing Jacquards over looms, 183 Placing looms in position in weaving sheds, 462 Plain side selvages, 445 side selvages, with boat, 445 side selvages, with mails, cam, and lever, 447 Points to be considered before con- structing a tappet, 29 Position of closed warp line with stocks and bowls, 75 of coupling knot, 187 of healds in loom, 7 of parts for beating up, 338 of preparatory machinery, 457 of slay when reed and fabric are in contact, 347 of slay swords to strike the fabric at right angles, 347 of tappet on loom, 25 of treadle fulcrum, 42 Positive beating up, 333 dobby, Christy's, 110 dobby, Knowles's, 113 drop boxes, 405 drop boxes, Shaw's, 405 let-off and take-up motions, Keighley's, 376 motion to shuttle, 276 pick, Lyall's, 305 picking, 305 taking-up motion, intermittent, 381 tappets, 47 Power of slay to beat up, 345 loom, introduction, 1 Pressure harness, 206 Principal motions of loom to be connected jDOsitively, 277 Principle of reversing motions, 62 Principles of dobby shedding, 80 of shedding, 18 Proportioning heald twine to reed and warp, 9 Pulling catch of taking-up motion, 384 Range of dobby shedding, 24 of heald shedding, 23 of Jacquard, 120 INDEX 471 Reading-ill and repeating card-cut- ting macliine combined, 232, 234 in card-cutting machine, 230 Reasons for slow development of power-loom, 1 Reed calculations, 351 Reed used in place of a comber board, 181 Reeds, 349 Reid, Fisher, and Parkinson's card- lacing machine, 248 Relative consumption of power by spring and weight, 65 consumption of power by stocks and bowls and weights, 69 Removal of rain water from roofs of weaving sheds, 460 Requirements of a box motion, 394 of a good letting- off motion, 365 Reversing motions, principles of, 62 Revolving boxes, 423 Rule for determining length of stroke to be given to a slay, 345 Salt's centre selvage motion, 442 Schilling's automatic warp weighting motion, 372 Scroll pick, revolving, 293 pick, stationary, 291 tappet, 61 Self-acting temples. Dr. Cartwright's, 427 Semi-open shedding, 22 Separate driving for dobby cylinder. 86 Setting a weft fork, 355, 456 of check strap, 312, 455 of loose reed motion, 324, 456 of pick, 453 of stop-rod blades, 321, 456 of takiiig-up motion, 383, 456 of temples, 433, 457 of weft fork tappet, 356, 456 Shakers, 221 Shape of picking tappet face, 301 Shaw's box motion, 405 Shedding, principles of, 18 Shields's Jacquard, 167 Shorrock and Taylor's centre selvage motion, 438 Short picking bands, 310 Shuttle boxes, timing of, 456 guard, automatic, Farrell's, 329 guard, semi-automatic. Hall and Sons', 327 guard, semi-automatic, Hamblet and Clifton's, 329 guard, semi-automatic, Marriott's, 328 guards, 326 guards, automatic, 329 guards, rigid, 326 boxes governed by shedding motion, 113, 403 Shuttles of equal weight and size, 277, 454 Singer Co.'s card-lacing machine, 249 Single-acting Jacquard, 122 acting reversing motions, 62 lift negative dobby, 81 picking bands, 310 roller temples, 430 roller temples for distending a fabric, 431 Sizes of driving wheels, 452, 463 Slay to swing from top framing of loom, 347 Slays with compound beats, 348 Smith Brothers' automatic letting- off motion, 374 Space between shaft centre and thin part of tappet, 35 Speeding of looms, 345 Split, or double scale harness, 203 Spring box, 129 easing motion, Hahlo, Liebreich, and Hanson's, 66 reversing motions, 63 tension cords for lappets, 259 Springs of equal strength, 455 Stocks and bowls, 68 Stop motion for double cylinder Jacquards, Devoge's, 151 motion for double cylinder Jac- quards, Green and Barker's, 149 rod, 320, 321 Straight tie harness, 193 Strain upon warp in shedding, 34 Strapping, 309 472 MECHANISM OF WEAVING Swell, Collier, Evans, and Riley's, 318 Harker and Grayson's, 325 Swells, fast reed, 318 Swivel shuttle, 315 shuttles moved by rack and pinion, 316 weaving, 313 weaving, process of, 314 weft, tension on, 315 Swivels and lappets compared, 313 Table showing movement of crank, 343 Taking-up motions, 380 motions governed by the reed, 391 Tappet, barrel, 45 driving, 58 Jamieson's, 43 Nuttall's chain, 57 oscillating, 54 ]3osition of, 25 positive, 47 shedding, 24 Woodcroft's, 49 Temples, 427 horizontal ring, 435 inclined ring, 433 single roller, 430 three roller, 432 two roller, 432 use of, 427 Tension on swivel weft, 315 Tie-ups, or cording plans, 29 Time of picking, 301, 453 Timing and fixing of parts of loom, 448 of shuttle boxes, 456 Top doup harness, 222 light in weaving sheds, 460 Treadle fulcrum, position of, 42 Trough and roller temples, Bul- lough's, 428 Turning back bar for Jacquard,143 Twilling Jacquard, 168 Twine loops, as doups, for centre selvages, 437 Under motion, Kenyon's, 67 picks, 278 Unequal compound harnesses, 197 Use of shakers, 221 Uses of a dobby, 24 of a Jacquard, 120 Value of a tappet for shedding, 24 Velocity of shuttle, varying, 273 Verdol's Jacquard, 164 Vertical punch, reading - in and card-repeating machine, 234 Ward's dobby, 102 Warp beam, 368 line, 450 Warped harness, 190 Warping harness threads, 184 Waste of power in picking, attempts to reduce, 275 Weaving rooms or sheds, 457 shed tioors, 462 shed roof, 460 shed walls, 457 Weft fork, and grid, 353 stop motions, 352 stop motion. Dr. Cartwright's, 353 Weight and rope governing motion for warp, 366 Whitesmith's box motion, 411 Width of loom framing, 449 of passages in weaving sheds, 459 Wilkinson's open-shed harness, 226 Wire healds, 10 Woodcroft's tappet, 49 tappet connection with healds, 53 tappet for open shedding, 51 tappet ring, 50 Yates's pick, 290 Printed by^ & R. Clark, Edinburgh. WILSON BROTHERS LIMITED, CORNHOLME MILLS, TODMORDEN. (ESTABLISHED 1823.) Largest and most complete Bobbin and Shuttle Works in the World. 18 Awards for Excellence of Manufacture. SHUTTLE DEPARTMENT. WILSON BROTHERS, Limited, make Shuttles in Boxwood, Cornel, Persimmon, Fruit-Tree, etc. The Blocks are all carefully selected and well seasoned. Every care is taken to produce goods of the best possible description, and embodying the latest improvements, so as to meet the varied requirements of our clients. Shuttle Tongues.— Wilson Brothers, Limited, have given special attention to the production of a Shuttle Peg as nearly perfect as possible. In order to overcome numerous difficulties they now make their own forgings. Every bar is tested before being forged, and only the best quality is used. A special machine has been designed by means of which pegs are being supplied as accurate in diameter as Ring Spindles. ACCESSORIES. Shuttle Pegs in Iron, Steel, or Brass. Shuttle Pegs neatly repaired. Shuttle Springs, Catches, and Wood Guards. Sluittle Pirns and Weavers' Bobbins in Boxwood, Beech, etc. Shuttle Pirns fitted with Holden's Patent Brass Tip, by which the Pirns are pre- vented from splitting and the tips are kept smooth. Shuttle Pins in Boxwood, Lignum Vitaj, or Ebony, also in Iron, Steel, or Brass. Picking Sticks and Picking Band Pegs. Warping Creel Pegs and Warping Creel Steps. Beam Wedges. PATENT WARPING OR WINDING BOBBINS. These Bobbins are fitted with our Patent Steel Binder, which is novel in construc- tion and is a perfect protection to the flanges. The Patent Binder makes it impossible for the flanges to open or come loose, and does not add materially to the weight of the Bobbin. It is applicable to Bobbins with flanges from 2 inches up to 18 inches diameter, and by its use there is a great saving of waste yarn through the prevention of broken flanges. Bobbins with these Patent Steel Binders have been in use for many years, and have met with unqualified success. ORIGINAL INVENTORS AND MAKERS OF STEEL AND BRASS-PLATED BOBBINS AND TUBES. We make a Speciality of Light Cardroom Tubes, with small diameters for fine counts. We have excellent facilities for cutting, seasoning, and partially preparing both Tubes and Bobbins before turning. We have large stocks of regular sizes always seasoning, the process occupying from two months up to two years. We are thus enabled to guarantee exact and uniform diameters. We are the largest makers of Ring Twist Bobbins and Ring Weft Pirns. Every Bobbin is carefully balanced. Ring Twist Bobbins, with Wilson Brothers' Patent " C " Shield, are the strongest Bobbins in the market, and are capable of resisting 150 per cent more pressure than plain Bobbins. Office and Showrooms (Open Daily): Branch Works : 14 Market Place, Manchester. Garston, Liverpool. LAMBETH COTTON ROPES. They are firmly made and very solid, containing more actual yarn for a given diameter than is usual ; and being made from pure Egyptian Throstle Yarn, without any weighting material, are light in weight. Also DRUM, RIM, SCROLL, SPINDLE, RING SPINDLE, TAPE, and TUBULAR BANDINGS to any description for Cotton Mills. THE LAMBETH COTTOIsr ROPES are of unique design and construction, superseding all other Cotton Ropes for Main Driving. Tension and Friction accurately measured for and provided against, and the Ropes fitted exactly to the working part of the grooves of the pulley. A LARGE STOCK OF ALL SIZES KEPT TO MEET URGENT ORDERS. NOTE.— These ropes are made at my works alone, and are only genuine when bearing my Registered Trade Marl(. THOMAS HART, BLACKBURN. Telephone No. 10. ESTABLISHED 1789. Telegraphic Address: ' (ABC Code used.) HART, BLACKBURN.' TECHNICAL HANDBOOKS. PEIMER of HYGIENE. By Ernest S. Eeynolds, M.D. Lond. Pott 8vo. Is. ELEMENTS of HANDICRAFT and DESIGN. By W. A. S. Benson, M.A. Oxon. With Illustrations. Extra Crown 8vo. 6s. net. " It is a solid, practical work — practical always with an intellectual as well as a merely educational side to it."—Sc]iool Board Chronicle. DRAWING and DESIGN. A Class Text-Book for Beginners. By E. R. TA.YLOK, Headmaster of the Birmingham Municipal School of Art. With Illustrations. Oblong Crown 8vo. 2s. ^d. "All the lessons are nicely calculated to cultivate eye, hand, and taste. . . . We have never seen a book of instruction for beginners comparable to this. It ought to be in the hands of every young person in the land." — Birmingham Daily Post. LEAD WORK. Old and Ornamental, and for the most part English. By W. R. Lethaby. With Illustrations. Extra Crown 8vo. 45. 6d. net. "We have in this little book of his the enunciation of ijrinciples of treatment and methods of working as exemplified in old examples of the craft." — British Architect. A HANDBOOK of GOLD-MILLING. By Henry Louis, Associate of the Royal School of Mines, Fellow of the Geological Society, etc. Crown 8vo. 10s. net. " It is a valuable companion, and should be found on the shelves, not only of the man in charge of the mill, but also of the student and mining engineer." — Mining Journal. WOOLLEN SPINNING. A Text-Book for Students in Technical Schools and Colleges, and for Skilful Practical Men in Woollen Mills. By Chakles Yickeeman, Author of " The Woollen Thread " and "Notes on Carding." With Illustrations. Crown 8vo. 6s. " To those who have some knowledge of the matter, we cannot recommend it too highly. It is certainly the most practical treatise which has been produced." — Bradford Observer. The GRAMMAR of WOODWORK. By Walter E. Degerdon, Head Instructor in Woodwork at the Whitechapel Craft-School. With a Preface by L. H. Llewellyn Smith. 4to. Paper Covers. 2s. "The order of work is clearly mapped out, and the book will be found to be most useful for reference at the bench. We warmly recommend it to the notice of all manual instructors. " — Schoolmaster. A TEXT -BOOK of NEEDLEWORK, KNITTING, and CiJTTiNG-ouT, with Methods of Teaching. By Elizabeth Rosevear. With Original Illustrations and Sectional Diagrams. Second Edition. Revised and Enlarged. Crown 8vo. 6s. "A very nseful book of reference for teachers of elementary sewing classes." — Saturday Review. NEEDLEWORK, KNITTING, and CUTTING-OUT, for Older Girls. By Elizabeth Rosevear. With Original Illustrations and Sectional Diagrams. Globe 8vo. Standard lY., 6d. ; Standard V., Sd. ; Standards VI., VII., and Ex-VIL, Is. NEEDLEWORK. By Elizabeth Glaister. Crown 8vo. 29. 6d. {Art at Home Series. MAOMILLAN AND CO., LONDON. TECHNICAL HANDBOOKS-Continued. CUTTING-OUT and DRESSMAKING. From the French of Mdlle. E. Grand'homme. With numerous Diagrams. Pott 8vo. Is, DRESSMAKING. By Mrs. Henry Grenfell. With Preface by Miss F. L. Calder. Pott 8vo. Is. A PRIMER of DOMESTIC ECONOMY. By Edith A. Bar- KETT and H. C. O'Neill. Pott 8vo. Is. A PRIMER of AGRICULTURAL CHEMISTRY ; or, the Food of Plants. By A. P. Laurie, M.A., Fellow of King's College, Cam- bridge ; Organising Secretary to the Bedfordshire County Council. Pott 8vo. l5. A PRIMER of PRACTICAL HORTICULTURE. By J. Wright, F.R.H.S. Pott 8vo. Is. By Prof. HENRY TANNEB, M.R.A.C., F.C.S., Examiner in the Principles of Agriculture under the Government Department of Science. The ALPHABET of the PRINCIPLES of AGRICULTURE. Extra Fcap. 8vo. 6d. FURTHER STEPS in the PRINCIPLES of AGRICULTURE. Extra Fcap. 8vo. Is. ELEMENTARY SCHOOL READINGS on the PRINCIPLES of Agriculture for the Third Stage. Extra Fcap. 8vo. Is. FIRST PRINCIPLES of AGRICULTURE. Pott 8vo. Is. ELEMENTARY LESSONS in the SCIENCE of AGRICUL- TURAL Practice. Fcap. 8vo. 3s. 6d. MACMILLAN'S COMMERCIAL CLASS-BOOKS. Edited by JAMES GOW, Litt.D., Headmaster of the High School, Nottingham. Globe 8vo. The HISTORY of COMMERCE in EUROPE. By H. DE B. GiBBiNS, M.A., Oxford University Prizeman in Political Economy, and Author of "The Industrial History of England." With Maps. 3s. 6d. COMMERCIAL ARITHMETIC. By S. Jackson, M.A., formerly Post Master of Merton College, Oxford ; lately Master in the Nottingham High School ; Headmaster of Victoria College, Congleton. 35. 6d. MACMILLAN AND CO., LONDON. Wr Date Due J - .